Patent Grant
10664169
This disclosure relates in general to the field of communications and, more particularly, to improving performance of object storage systems.
Data storage is a primary function performed by even the most rudimentary computing systems. Data is often stored in binary form (e.g., a string of bits, each of which is either a zero or a one). Binary data can be encoded by using a particular pattern to correspond to individual alphanumeric characters, media, and other digital groupings. However, since humans cannot easily read stored binary data or encoded data, the data is grouped into files, each of which is given a human-readable name. Files are managed by a file system. There exist myriad file storage systems for storing the files upon which the file system is built. For example, some file storage systems directly store files in a local storage disk while others distribute files to one or more remote storage disks. In object storage systems, each file is split into several portions, called objects, before being stored.
To provide a more complete understanding of the present disclosure and features and advantages thereof, reference is made to the following description, taken in conjunction with the accompanying figures, wherein like reference numerals represent like parts, in which:
Some file storage systems store each complete and unmodified file in a contiguous block of memory in a storage disk. However, some object storage systems use a process, sometimes referred to as data striping, to split each file into several portions, called objects, before storing them in one or more storage devices. In other words, each object is a fragment of a file (e.g., each fragment being a subdivision of the file). It is noted that the terms ‘object’ and ‘fragment’ are used interchangeable in the present disclosure. In addition, each object may be replicated on more than one storage device. For example, when writing a file, the file is split into objects and each object may be stored on a storage device (a primary storage device) and a copy of the object may be stored on, e.g., one or more other storage devices (redundant storage devices referred to herein as replica storage devices). It is noted that each object of a single file need not have the same storage device as the primary storage device. Indeed, the objects that comprise a single file may be distributed and replicated across 100s of storage devices. When the file is read from the file system, each object is read only from the primary storage device. Thus, reading and writing an object is asymmetric in that reading the object from an object storage system (OSS) only requires accessing a single storage device (i.e., the primary storage device) while writing an object to the OSS requires accessing a several storage devices (i.e., the primary storage device and each of the replica storage devices). In many systems, it is desirable to keep latency down (e.g., low latency is preferred; high latency is undesirable). In general, this asymmetry causes objects to be read faster (e.g., lower latency) than they are written. Reading and writing files causes repeated reading and writing of objects, which further exacerbates the asymmetry.
An object storage system (OSS) is inclusive of a plurality of object storage devices (OSDs) for storing objects (i.e., the file fragments). An object storage device (OSD) is a storage device (e.g., physical or logical) in which the objects are stored. In many cases, an OSD is a physical storage device (including, e.g., a memory, a processor, a cache) but may also be inclusive of a logical disk (based on several physical devices or partitions of a single physical device). An OSD is a primary storage unit of an OSS. An OSD can include a client (e.g., code, software) that can be executed by an endpoint to access a logical disk using a file system interface. A ‘logical disk’ is inclusive of a virtual storage device that provides access to memory that is located on one or more physical storage devices. For example, a logical disk may include memory from several different physical devices, each of which may be co-located (e.g., in a single data center) or may be remote from another (e.g., in different data centers). In addition, each physical device may have multiple partitions of memory each of which are can be used, by itself, as a logical disk or can be used in combination with memory from other storage devices to form the logical disk.
One problem that arises in file systems is how to determine the latency of a disk on which the files are stored. Computing latency is trivial for file systems that store files only on a local storage disk. Moreover, such calculations may be unnecessary since the disk is local and the files can be stored in a contiguous block of memory. For file systems that store (all) files of the logical disk in a single remote disk, the latency of the logical disk is simply the latency of the single remote disk. For file systems that store files in multiple remote disks, latency can be determined based on a latency of each remote disk and a weight value that accounts for a proportion of the logical disk stored in each remote disk. For example, the logical disk may store files in two remote disks, where 20 percent of the files of are stored in a first disk and 80 percent of the files are stored in a second disk. In such a case, the latency of the logical disk may be calculated as 0.2*(latency of the first disk)+0.8*(latency of the second disk). However, it is much more complicated to determine the latency of a logical disk for which the data is stored in an object storage system (e.g., based on the asymmetry of the reads/writes of objects) and the large number of storage devices that may be used for a single file. Thus, an objective technical problem is to determine the latency of a logical disk that asymmetrically reads object from and writes objects to the logical disk.
A file system may utilize an object storage system (OSS) as an underlying data storage system. For example, a service provider (or operator) may provide data storage as a service to endpoints using an OSS. Each endpoint may be associated with an entity (e.g., an individual or an organization). A tenant refers an entity associated with one or more endpoints each of which can receive, as a service from the service provider, access to the file system and/or the OSS. Each endpoint is associated with (e.g., belongs to and/or is operated by) at least one entity. As an example, a company (i.e., an organization that is a tenant of the service provider) may provide each of its employees with a mobile device (i.e., an endpoint). Each endpoint may acquire (e.g., download), from the service provider, a client module (e.g., code, software), which, when executed by the endpoint, generates a file system interface based on objects stored in the OSS. Thus, from the perspective of any endpoint that uses the client module and/or the file system interface to access the file system, the (complete and unmodified) files appear to be stored in a single directory. However, on the backend of the file system, each file is fragmented into objects and is stored across one or more storage devices.
The term ‘endpoint’ is inclusive of devices used to initiate a communication, such as a computer, a personal digital assistant (PDA), a laptop or electronic notebook, a cellular telephone (e.g., an IPHONE, an IP phone, a BLACKBERRY, a GOOGLE DROID), a tablet (e.g., an IPAD), or any other device, component, element, network element, or object capable of initiating voice, audio, video, media, and/or data exchanges within system 100 (described below). An endpoint may also be inclusive of a suitable interface to the human user, such as a microphone, a display, or a keyboard or other terminal equipment. An endpoint may also be any device that seeks to initiate a communication on behalf of another entity or element, such as a program, a conferencing device, a database, or any other component, device, element, or object capable of initiating an exchange within the system 100. Furthermore, endpoints can be associated with individuals, clients, customers, or end users. Data, as used herein in this document, refers to any type of numeric, voice, messages, video, media, or script data, or any type of source or object code, or any other suitable information in any appropriate format that may be communicated from one point to another.
The network 118 operatively couples the tenants (i.e., 102 and 106) and the data center 124 to one another. The network 118 facilitates two-way communication between any two or more of the components of system 100. For example, each of the endpoints 104 can transmit to and/or receive data from the data center 124 (e.g., the network element 124, logical disks 128 and 138, and/or storage devices 132a-e therein) over the network 118. Within the context of the disclosure, a ‘network’ represents a series of points, nodes, or network elements of interconnected communication paths for receiving and transmitting packets of information that propagate through a communication system. A network offers communicative interface between sources and/or hosts, and may be any local area network (LAN), wireless local area network (WLAN), metropolitan area network (MAN), Intranet, Extranet, Internet, WAN, virtual private network (VPN), or any other appropriate architecture or system that facilitates communications in a network environment depending on the network topology. A network can comprise any number of hardware or software elements coupled to (and in communication with) each other through a communications medium.
The data center 124 comprises a network element 126 and a plurality of storage devices 132a-132e. The network element 126 and each of the plurality of storage devices 132a-132e are operably coupled to one another by communication channels. The data center 124 includes two logical disks: logical disk 128 and logical disk 130. The logical disks 128 and 130 are associated with at with tenants 120 and 106 respectively. Each tenant is provided with access only to its logical disks (and not to logical disks that are associated with other tenants). The network element 126 may maintain metadata associated with the logical disks. In this example, each tenant is associated with a single logical disk only for clarity of the figures. Each tenant may be associated with any number of logical disks. The metadata may include data mappings that, at least in part, define the logical disks and identify a manner in which the logical disk operates. For example, the network element 126 can store data comprising: a mapping of each storage device to one or more logical disks, a mapping of each tenant to one or more logical disks; a mapping of each tenant to one or more storage devices; a mapping of each logical disk to operational variables that identify a manner in which a logical disk operates. The operational variable may include (but are not limited to): a size of each object to be stored in the logical disk (e.g., measured in bit or multiples thereof), a number of primary storage devices in which each object is to be stored, a number of replica storage devices in which each object is to be stored, pools of storage devices (e.g., subgroups of storage devices within a logical disk) for which one of the storage devices is primary storage device and others of the storage devices are replica storage devices, and/or any other parameters that define (or otherwise specify) a manner in which a logical disk operates. In this example, each of the storage devices 132a, 132b, and 132c is mapped to logical disk 128; each of the storage devices 132c, 132d, and 132e is mapped to logical disk 130; the tenant 102 is mapped to the logical disk 128; the tenant 106 is mapped to the logical disk 130; the logical disk 128 is associated with operational variables including: a fragment size of three bits, one primary storage device for each fragment, two replica storage devices for each fragment; and the logical disk 130 is associated with operational variables including: a fragment size of four bits, one primary storage device for each fragment, and one replica storage device for each fragment. It is noted that each logical disk may be mapped to a single logical disk or to multiple logical disks. In this example, the logical disks 128 and 130 share the storage device 132c.
As used herein in this Specification, the term ‘network element’ is meant to encompass any as servers (physical or virtual), end user devices, routers, switches, cable boxes, gateways, bridges, load balancers, firewalls, inline service nodes, proxies, processors, modules, or any other suitable device, component, element, proprietary appliance, or object operable to exchange, receive, and/or transmit data in a network environment. These network elements may include any suitable hardware, software, components, modules, interfaces, or objects that facilitate the sharing of message queue operations thereof. This may be inclusive of appropriate algorithms and communication protocols that allow for the effective exchange of data or information. Each of the network elements can also include suitable network interfaces for receiving, transmitting, and/or otherwise communicating data or information in a network environment.
In one particular instance, the architecture of the present disclosure can be associated with a service provider deployment. For example, a service provider (e.g., operating the data center 124) may provide the tenants 102 and 106 with access to logical disks 128 and 130 in the data center 124. In other examples, the architecture of the present disclosure would be equally applicable to other communication environments, such as an enterprise wide area network (WAN) deployment. The architecture of the present disclosure may include a configuration capable of transmission control protocol/internet protocol (TCP/IP) communications for the transmission and/or reception of packets in a network.
The dashed lines between the network 118 and the tenants 102 and 106 (and the endpoints 104 therein) and the data center 124 represent communication channels. As used herein, a ‘communication channel’ encompasses a physical transmission medium (e.g., a wire) or a logical connection (e.g., a radio channel) used to convey information signals (e.g., data, data packets, control packets, messages etc.) from one or more senders (e.g., an tenant, an endpoint, a network element, a storage device, and the like) to one or more receivers (e.g., a second data center, a second message queue, a message consumer, a network element, and the like). Data, as used herein, refers to any type of source or object code, object, fragment of a file, data structure, any type of numeric, voice, messages, video, media, or script data packet, or any other suitable information in any appropriate format that may be communicated from one point to another. A communication channel, as used herein, can include one or more communication links, which may be physical (e.g., wire) or logical (e.g., data link, wireless link, etc.). Termination points of communication channels can include network interfaces such as Ethernet ports, serial ports, etc. In some examples, each communication channel may be a single channel: deployed for both control messages (i.e., instructions to control a network element, a logical disk, and/or a storage device) and data messages (i.e., messages that include objects for storage in one or more storage devices).
The file system interfaces 110 and 112 are a graphical user interface for the file system that stores files in logical disks 128 and 130 (respectively) in the data center 124. Each of the endpoints 104 executes a code block to generate the file system Interface (i.e., interface 110 or 112) through which the endpoint can access the files in the file system. Each of the tenants is provided with access to one or more logical disks in the data center. Tenants have a corresponding customized file system interface through which to access to their logical disks (e.g., the interfaces are customized on a per-tenant basis). The file system interface 110 renders filenames (i.e., filenames “FILE 1”, “FILE 2”, . . . , “FILE n”) for a directory location in the logical disk 128 associated with the tenant 102. The file system Interface 112 renders filenames (i.e., filenames “FILE A”, “FILE B”, . . . , “FILE m”) for a directory location in the logical disk 130 associated with the tenant 106. Each of the filenames corresponds to a file. The filename 114 corresponds to file 116. The filename 120 corresponds to file 122. Within the file system interfaces 110 and 112, files appear to be complete, unmodified, and stored in a single directory. Thus, from the perspective of any of the endpoints 104 that use the file system interface, the (complete and unmodified) files appear to be stored in a single directory. However, within the data center 124, each file is fragmented into objects and is stored across one or more storage devices.
For example, each of the files 116 and 122 is fragmented into objects; the objects are stored distributed and replicated across several storage devices in a logical disk. The file 116 is named “FILE 1” and corresponds to the filename 114 in the file system interface 110. The file 122 is named “FILE A” and corresponds to the filename 120 in the file system interface 112. Each of the files 116 and 122 is stored, at least in part, in binary form. The files may be further encoded in a standardized encoding (e.g., ASCII, Unicode, BinHex, Uuencode, Multipurpose Internet Mail Extensions (MIME), multimedia encoding such as audio and/or video encoding) or a propriety encoding (e.g., a proprietary file type generated by proprietary software). The files are split into objects. Each object is a fragment of the file. A size of the objects (e.g., measured in bit, bytes, or any multiple thereof) is configurable (e.g., by an endpoint, a network element, and/or other computing element with administrative rights). Thus, each service provider, each tenant, each endpoint, and the like may customize the size of the object. In the example of system 100, one of the endpoints 104a-d has set the size of the object to 3 bits for all files associated with the tenant 102; one of the endpoints 104e-i has set the size of the object to 4 bits for all files associated with the tenant 106. The files 116 and 122 are split into objects based, at least in part, on the size of the object set for the corresponding tenant. The file 116 (i.e., “FILE 1”) is split into five 3-bit objects (labeled “F1.1”, “F1.2”, “F1.3”, “F1.4”, and “F1.5”). The file 122 (i.e., “FILE A”) is split into three 4-bit objects (labeled “FA.1”, “FA.2”, and “FA.3”). Each object (i.e., each fragment of the file) is asymmetrically stored in and retrieved from storage devices in the data center 124.
Each fragment of a file is asymmetrically stored in and retrieved from storage devices in the data center 124. Each fragment being stored in multiple storage devices makes the logical disk robust. A failure of any of the storage devices is less likely to cause a complete loss of any file or fragment thereof at least because each fragment is redundantly stored in multiples storage devices. If a storage device fails, copies of fragments stored on others of the storage devices can be used to reassign a primary and/or additional replica storage devices as needed. Each fragment is assigned a primary storage device and one or more replica storage devices from the storage devices associated with a logical disk. A primary storage device stores its assigned fragments and, when they are requested from the logical disk, the primary storage device retrieves the assigned fragments and transmits them to the requesting component. A replica storage device stores its assigned fragments and but is not responsible for retrieving the assigned fragments in response to requests. In other words, while the each fragment is stored in multiple storage devices (i.e., the primary and the replica storage devices) only one of the multiple storage devices responds to requests to read the fragment. Writing a fragment to the logical disk requires activity from several storage devices (i.e., the primary storage device and each of the replica storage devices) while reading the fragment from the logical disk only requires activity from the a single storage device (i.e., the primary storage device). Retrieving an object from the logical drive accesses the storage devices a different number of times than for writing the object to the logical drive.
Each object is stored in a primary storage device and one or more replica storage devices. In
The file 122 (i.e., “FILE A”) is split into three 4-bit objects (labeled “FA.1”, “FA.2”, and “FA.3”) and stored in the logical disk 130. The logical disk 130 comprises storage disks 132d and 132e and at least a portion of the storage disk 132c. Each object of File A is stored in a primary storage device and a replica storage device. For example, for the object FA.1, the storage device 132c is the primary storage device (storing the object FA.1); and the storage device 132e is a first replica storage device (storing the first copy FA.1′). Table 2 below summarizes the primary storage devices and replica storage devices for each of the objects parsed from the FILE A.
As discussed above, an objective technical problem is to determine the latency of a logical disk that asymmetrically reads and writes objects, as is the case for logical disks 128 and 130. The following is a practical example of such a problem. In a cloud computing environment (e.g., the cloud computing software marked under the trade name OpenStack) with an object storage system (e.g., the storage system marketed under the trade name CEPH) as the backend storage, it is common for an operator (e.g., a service provider) to have difficulties identifying a reason why a tenant's logical disk (e.g., a Ceph volume) is “slow” (i.e., read and/or write operations to the logical disk have a latency that negatively impacts performance of the system and/or is unacceptable to the tenant). When a tenant reports, to the operator, that their logical disk is slow, operators are neither able to validate the report nor identify the reasons for the logical disk being slow. Because the logical disk volume is distributed across potentially thousands of physical disks (e.g., storage devices), isolating the problem to one or more disks can be a challenge.
A potential solution is to empirically determine the latency of a disk. In traditional distributed storage systems, latency can be calculated through experiments based on request and response (completion) of events. This method of latency calculation cannot be applied to determine latency of a logical disk (such as logical disks 128 and 130 of
A solution, disclosed in the present disclosure, to address the above issues (and others) provides for improving performance of a logical disk by reconfiguring storage devices in the logical disk based on asymmetric reading and writing characteristics of the storage devices. The methods, systems, logic, and/or apparatuses (as disclosed herein) address the above technical problem (and others) by adding and/or removing storage devices from the logical disk based on an influence that each of the exiting storage devices in the logical disk has on the overall performance of the logical disk. In some examples, the methods, systems, logic, and/or apparatuses disclosed herein utilize a number of objects that a storage device is associated with retrieving and a different number of objects that the storage device is associated with storing to determine the influence on the logical disk. In addition, the adding or removing of the storage device can be simulated using a mathematical model of the logical disk to verify whether the addition or removal (as the case may be) of the storage device will improve the performance of the logical disk.
The network element 126 of
Each of the storage devices 132a-c comprises respective processors 216a-c, network interfaces 218a-c, and memory elements 222a-c. Each of the network interfaces 218a-c includes a respective plurality of ports 220a-c, each of which is configured to transmit and/or receive data over a network. Each of the memory elements 222a-c stores, among other things, a distributed storage code block 226a-c. The processors 216 execute the distributed storage code blocks 226a, which, at least in part, define and manage an object storage system in the data center 124. When executed, each of the distributed storage code blocks 216a-c, can generate control plane messages for communication with corresponding distributed storage code blocks in other components of the logical disk (e.g., distributed storage code block 207 in the network element 126). Moreover, each distributed storage code block the logical disk may include libraries of functions to operate the OSS (e.g., algorithms for distributing objects to storage devices and/or copying objects to replica storage devices). Each of the storage devices 132a-c has a portion of their memory element dedicated to storing objects. Storage device 132a include memory portion 224a, which stores objects (and copies of objects) associated with files stored in the logical disk 128 of
The table 300 includes metadata corresponding to the storage devices 132a-e and the logical disks 128 and 130 in the data center 124 of
The row 312 corresponds to metadata associated with the storage device ID 1 (i.e., the storage device 132a) within the context of the logical disk ID 1 (i.e., the logical disk 128). The row 314 corresponds to metadata associated with the storage device ID 2 (i.e., the storage device 132b) within the context of the logical disk ID 1 (i.e., the logical disk 128). The row 316 corresponds to metadata associated with the storage device ID 3 (i.e., the storage device 132c) within the context of the logical disk ID 1 (i.e., the logical disk 128). The row 318 corresponds to metadata associated with the storage device ID 3 (i.e., the storage device 132c) within the context of the logical disk ID 2 (i.e., the logical disk 130). The row 320 corresponds to metadata associated with the storage device ID 4 (i.e., the storage device 132d) within the context of the logical disk ID 2 (i.e., the logical disk 130). The row 322 corresponds to metadata associated with the storage device ID 5 (i.e., the storage device 132e) within the context of the logical disk ID 2 (i.e., the logical disk 130).
Each of the rows of the table 300 identifies a combination of a logical disk ID and a storage device ID and corresponding performance parameters for the combination. The rows identify the combination of the logical disk ID and the storage device ID at least because each storage device may be associated with more than one logical disk. For example, the storage device ID 3 (i.e., the storage device 132c) is associated with both the logical disk IDs 1 and 2 (i.e., the logical disks 128 and 130, respectively) and, as a result, the table 300 contains two rows (i.e., rows 316 and 318) that identify the metadata for the storage device ID 3: one for each of the logical disk IDs 1 and 2.
Each impact factor identifies, on a per logical disk basis, a proportion of objects stored in or retrieved from each of the plurality of storage devices relative to others of the plurality of storage devices based on asymmetrical storage in and retrieval from each of the plurality of storage devices. The impact factors correspond to an influence that each of the storage devices has on the performance (e.g., average latency, amount of throughput, and the like) of the logical disks to which the storage device is associated. For example, as the amount of throughput (e.g., a computational load) on each storage device increases, the latency of each storage device increases because it takes more the time to process each read/write request than if the amount of throughput were reduced.
A network element (e.g., a server, such as a metadata server in a CEPH system), may calculate the impact factors by, at least in part, counting a number of objects for which each of the plurality of storage devices is a primary storage device and/or a replica storage device.
The read impact factor (i.e., in column 306) identifies a number of objects for which each of the storage devices is a primary storage device within the context of a logical disk. The network element may calculate the read impact factor based on metadata associated with the logical disk and/or data retrieved from the storage devices. For example, the network element may utilize metadata including the data of Tables 1 and 2 of the present disclosure to count the number of objects for which each of the storage devices is a primary storage device within the context of a logical disk. In other examples, the network element may transmit to each storage device in a logical disk a request for a read impact factor. The request for the read impact factor may be for a single file, multiple files, or for all files for which the storage device stores objects (based on identifiers of files and/or objects in the request). The request may also identify a particular logical disk for which the impact factor is requested (e.g., since the impact factor may be different for each logical disk to which the storage device is associated).
Assuming, only for the sake of an simple example, that each logical disk only stores a single file (e.g., FILE 1 or File A), the read impact factor for each device is the number times that the storage device is identified in the column labeled “Primary” in Tables 1 and 2 (i.e., the number of objects for which the device is the primary storage device, which responds to read requests for the object). Table 1 includes metadata for the logical disk 128 (logical disk ID 1 in
The above process is described with respect to calculating the read impact factor of various storage devices for a single file. Object storage systems often store many files. The impact factor for all of the files in the logical disk may be calculating by repeating, for each file in the logical disk, the above-described process of calculating impact factors for a single file. The overall read impact factor for each storage devices in the logical disk may be calculated by summing the individual read impact factors (for each file) for each storage device to determine.
The write impact factor (i.e., in column 308) identifies a sum of: a first number of objects for which each of the storage devices is a primary storage device, and a second number of objects for which each of the storage devices is a replica storage device. A network element may calculate the write impact factor based on metadata associated with the logical disk and/or data retrieved from the storage devices. For example, the network element may utilize metadata including the data of Tables 1 and 2 of the present disclosure to count the number of objects for which each of the storage devices is a primary storage device and the number of objects for which each of the storage devices is a replica storage device within the context of a logical disk. In other examples, the network element may transmit to each storage device in a logical disk a request for a write impact factor. The request for the write impact factor may be for a single file, multiple files, or for all files for which the storage device stores objects (based on identifiers of files and/or objects in the request). The request may also identify a particular logical disk for which the impact factor is requested (e.g., since the impact factor may be different for each logical disk to which the storage device is associated).
The following example assumes (only for the sake of a simple example) that each logical disk only stores a single file (e.g., FILE 1 or File A). In such an example, the write impact factor for each device is the sum of (1) a number times that the storage device is identified in the column labeled “Primary” in Tables 1 and 2 (i.e., the number of objects for which the device is the primary storage device), and (2) a number times that the storage device is identified in any of the remaining “Replica” columns in Tables 1 and 2 (i.e., the number of objects for which the device is a replica storage device, which responds to write requests for the object). Table 1 includes metadata for the logical disk 128 (logical disk ID 1 in
The impact factors are attributes that can be used to determine the influence that each storage device has on a logical disk by accounting for asymmetric read and write operations of the storage device. Other attributes may be used to determine the influence that each storage device has on a logical disk. Other attributes may include (but are not limited to) any one or more of the following attributes of a storage device: a number of pending operations (e.g., a number of operations in a queue of operations to be performed by the storage device), a number of stored objects (e.g., a total number of objects stored by the storage device across all of the logical disks with which it is associated), total number of logical disks using the storage device, and/or system metrics (e.g., performance parameters, a latency of input/output (“I/O”) operations measured during a window of time, current processor utilization, current memory utilization, and the like). For example, as the current utilization of the processor (e.g., percent of processor capacity utilized by currently executing processes) increases for each storage device, the influence that each storage device has on latency of the logical disk increases because it takes more the time for each storage device to process each read/write request than if the current utilization were reduced.
The attributes may be requested directly from each storage device or may be calculated. In the example of latency, a latency of I/O operations performed by a storage device may be calculated by dividing a number of operations performed by the storage device (during in a window of time) divided by the length of time interval (e.g., measured seconds, minutes, or multiples thereof) to get average latency of the I/O operations. In some examples, the window of time is a moving time window (e.g., a most recent window of 15 minutes, 30 minutes, 1 hour, and the like).
After the attributes are collected for (e.g., calculated and/or retrieved from) each of the storage devices in a logical disk known, a numerical representation of an influence that each storage device has on operational performance the logical disk is determined. The numerical representation corresponds to a proportion of the operational performance of the logical device that is attributable to each of the plurality of storage devices based, at least in part, on the attributes of each of the storage devices. The numerical representation of the influence of each storage device may be determined using a mathematical model. In some examples, the numerical representation is a weighting factor for each of the plurality of storage devices.
A weighting factor for each of the storage devices can be calculated based one or more of the attributes of the each of the storage devices. The weighing factors may be percentage values that correspond to the asymmetric reading and writing characteristics of the storage devices of the logical disk. A weighting factor for a storage device is determined, at least in part, based on an impact factor of the storage device. In some examples, the weighting factor is determined based only on the impact factor(s). The following illustrates example calculations for determining the weighting factors is determined based only on impact factors. Weighting factors are calculated in the context of a logical disk (e.g., on a per-logical disk basis). Returning to table 300 of
In other examples, the weighting factor is determined based on the impact factor(s) in combination with one or more other attributes. For example, a mathematical algorithm (e.g., regression algorithm performing a regression analysis) may be executed on impact factor in combination with one or more other attributes to determine weighting factors for the storage device. As a further example, a simulation may be executed be under various conditions of storage disks (e.g., various pseudo-random distributions of objects across the of storage disks, various numbers of pending operations, and various system metrics). A (theoretical) latency of the logical disk can be determined from such simulations. A regression algorithm may take as input: (1) independent variables or factors (e.g., impact factor(s), number of pending operations, number of stored objects, system metrics, and the like) and dependent variables such as a latency of a logical disk (i.e., determined based on the simulations). The regression algorithm may generate, as output, weights to satisfy the following equation for latency of a logical disk (LDLatency):
LDLatency=sum[WF(i)*LS(i)], for i=1, 2, . . . , n; where n is the number of storage devices in the Logical Disk
Where LS(i) and the latency of the ith storage device;
WF is the weighting factor for the storage device;
WF=sum[x(j)*f(j)], for j=1, 2, . . . , m; where m is the number of factors input for the storage devices; and
f(j) is the jth factor for the storage device (e.g., factors such as impact factor(s), number of pending operations, number of stored objects, system metrics, and the like)
x(j) is the jth weight associated with the jth factor.
The regression algorithm is utilized to generate the weights (i.e., x(j)). Once the weights are known, they are used to determine LDLatency in a production environment (for actual storage devices and not simulated storage devices). The mathematical algorithm can be executed on the fly (e.g., in near real-time) and at regular intervals to track the influence of a storage device over time.
The above examples of calculating the weighting factors are provided for illustration purposes only (and do not limit the teaching of the present disclosure). Indeed, the teachings of the present disclosure are equally applicable to any approach of calculating the weighting factors as long as the approach accounts for the asymmetric reading and writing characteristics of the storage devices of the logical disk, as is demonstrated by the above exemplary approaches.
A portion of a performance parameter of a logical device that is attributable to each storage device in the logical device is calculated based, at least in part, on a corresponding performance parameter of each storage device weighted by the corresponding influence. In the example of latency, a weighted latency for a storage device can be calculated based on latency of I/O operations of the storage device and the weighting factor for the storage device. The weighted latency of the storage device is a portion of latency of the logical disks that is attributable to the storage device. For example, the weighted latency of the storage device (WLatency) is calculated by multiplied the weighting factor for the storage device (WF) and the latency of the storage device (LS) (i.e., WLatency=WF*LS). In some examples, the latency of the storage device (LS) is the latency of the I/O operations measured during a window of time. A calculation can be performed for each of a plurality of plurality of storage devices comprising logical disk. In such an example, for each of the plurality of storage devices, the latency of the I/O operations is multiplied by the weighting factor to determine a weighted latency for each of the plurality of storage devices. Table 300 (in
A performance parameter of a logical disk can be calculated based, at least in part, on corresponding performance parameters each of the storage devices and the weighting factor. Thus, the performance parameter of the logical disk can be calculated by summing the weighted performance parameter for each of the plurality of storage devices in the logical disk. In one example, a latency of the logical disk is calculated by summing the weighted latency for each of the plurality of storage devices in the logical disk. Thus, the latency of a logical disk=sum[W(i)*Latency(i)] (for i=1, 2, . . . , n; where n is the number of storage devices in the Logical Disk). For example, the logical disk ID 1 (rows 312, 314, and 316 of
A proportion of the performance parameter of a logical device that is attributable to each of the plurality of storage devices is calculated based, at least in part, an impact factor and a latency of each of the plurality of storage devices. For example, the proportions can be determined using the calculations discussed with respect to tables 3 and 4, which are based on the impact factors and latencies of table 300 of
When the influence of each of the storage devices on a performance parameter of the logical disk is known, the performance of the logical disk is improved by reconfiguring one or more of the plurality of storage disks based on the influences. For example, a network device or endpoint may automatically (e.g., without any further input or prompting) toggle a storage device on or off and, thereby, improves the overall performance (e.g., reduces latency) of the logical disk.
The performance parameters of each storage device can have an unexpected impact on the performance parameter of the logical disk due to the objects being asymmetrically stored in and retrieved from the storage devices (and/or being pseudo-randomly distributed across the storage devices). In the example of table 300 (
In one implementation, the network elements, endpoints, servers, and/or storage devices, described herein may include software to achieve (or to foster) the functions discussed herein for improving performance of logical disks where the software is executed on one or more processors to carry out the functions. This could include the implementation of instances of file system clients, logical disk metadata, logical disk management modules, distributed storage code blocks, and/or any other suitable element that would foster the activities discussed herein. Additionally, each of these elements can have an internal structure (e.g., a processor, a memory element, etc.) to facilitate some of the operations described herein. In other embodiments, these functions for improving performance of logical disks may be executed externally to these elements, or included in some other network element to achieve the intended functionality. Alternatively, network elements, endpoints, servers, and/or storage devices may include software (or reciprocating software) that can coordinate with other network elements/endpoints in order to achieve the performance improvement functions described herein. In still other embodiments, one or several devices may include any suitable algorithms, hardware, software, components, modules, interfaces, or objects that facilitate the operations thereof.
In certain example implementations, the performance improvement functions outlined herein may be implemented by logic encoded in one or more non-transitory, tangible media (e.g., embedded logic provided in an application specific integrated circuit [ASIC], digital signal processor [DSP] instructions, software [potentially inclusive of object code and source code] to be executed by one or more processors, or other similar machine, etc.). In some of these instances, one or more memory elements can store data used for the operations described herein. This includes the memory element being able to store instructions (e.g., software, code, etc.) that are executed to carry out the activities described in this Specification. The memory element is further configured to store databases such as mapping databases (mapping various aspects of a logical disk to storage devices, clients, or other metadata) to enable performance improvements of a logical disk as disclosed herein. The processor can execute any type of instructions associated with the data to achieve the operations detailed herein in this Specification. In one example, the processor could transform an element or an article (e.g., data) from one state or thing to another state or thing. In another example, the activities outlined herein may be implemented with fixed logic or programmable logic (e.g., software/computer instructions executed by the processor) and the elements identified herein could be some type of a programmable processor, programmable digital logic (e.g., a field programmable gate array [FPGA], an erasable programmable read only memory (EPROM), an electrically erasable programmable ROM (EEPROM)) or an ASIC that includes digital logic, software, code, electronic instructions, or any suitable combination thereof.
Any of the devices disclosed herein (e.g., the network elements, endpoints, servers, storage devices, etc.) can include memory elements for storing information to be used in achieving the performance improvements, as outlined herein. Additionally, each of these devices may include a processor that can execute software or an algorithm to perform the activities as discussed in this Specification. These devices may further keep information in any suitable memory element [random access memory (RAM), ROM, EPROM, EEPROM, ASIC, etc.], software, hardware, or in any other suitable component, device, element, or object where appropriate and based on particular needs. Any of the memory items discussed herein should be construed as being encompassed within the broad term ‘memory element.’ Similarly, any of the potential processing elements, modules, and machines described in this Specification should be construed as being encompassed within the broad term ‘processor.’ Each of the devices can also include suitable interfaces for receiving, transmitting, and/or otherwise communicating data or information in a network environment.
At a high level, the logic 500, when executed, improves the performance of a logical disk. Logic 500 may be implemented in a network element 126 (of
In some embodiments, prior to a storage device being removed from and/or added to the logical disk, any reconfiguration of the logical disk may be simulated. The simulation may utilize a mathematical modeling of the performance of the logical disk to estimate the affect of the reconfiguration on the logical disk. The mathematical model can take, as input, attributes of the storage disks that comprise the logical disk and generate, as output, a performance parameters of the logical disk (including a breakdown of the proportion of the performance parameter attributable to each storage disk). Thus, the mathematical model can be used to determine current performance parameters of the logical disk and simulated performance parameters of the logical disk after the reconfiguring of the logical disk (by simulating objects being redistributed to the reconfigured storage devices). By comparing the current performance parameters to the simulated performance parameters, the behavior of the logical disk can be assessed to assess whether the reconfiguration will improve the performance of the logical disk (e.g., improve the performance by reducing latency). Reconfiguring the logical disk by adding or removing storage devices Ultimately results in the objects of the being redistributed. In the case of adding a storage device, objects are removed from other storage device and added to the new storage devices (e.g., pseudo-randomly selected for reassignment to the new storage device using an algorithm such as CRUSH). During the simulation, the CRUSH algorithm may be used to simulate assigning the objects to a new storage device by only determining a new location for the objects (however, the objects are not actually relocated). When it is determined, based on the simulation, that the reconfiguration will improve the performance of the logical disk, the reconfiguration is implemented (e.g., by adding or removing a storage device and actually relocating the objects). When it is determined, based on the simulation, that the reconfiguration will not improve the performance of the logical disk, the reconfiguration is not implemented. For example, the latency of the logical disk can be estimated using new impact factors derived from a simulated reassignment of objects (e.g., from a storage device that is to be removed from the logical disk). If it is determined, based on the simulation, that the storage device being turned-off meets a performance benchmarks (e.g., reduces latency below a threshold), the storage device is removed from the logical disk. If it is determined, based on the simulation, that the storage device being turned-off does not meet the performance benchmarks (e.g., does not reduce latency below the threshold), the storage device is not removed from the logical disk. Instead, a storage device may be added to the logical disk. Again, before such a reconfiguration is implemented, it may be simulated to assess whether the change would improve the performance of the logical disk.
At 612, at least one of the plurality of storage devices is removed from the logical disk. In embodiments where simulation is used, the simulation may have determined that removing the at least one of the plurality of storage devices improves the performance (performance parameter) of the logical disk. To remove a storage device from the logical disk, the objects on the storage device are relocated from the storage device to others of the plurality of storage devices. For example, all objects associated with files stored in the logical disk may be distributed (e.g., pseudo-randomly or with a distribution skewed in favor of storage devices with the best performance parameters) from the storage device to the others of the storage devices (e.g., using the CRUSH algorithm, or any other algorithm operable to determine locations for the objects). After all the objects associated with files stored in the logical disk are copied to the new locations, they are deleted from the storage device. In addition, the storage device is disassociating from the logical disk. This disassociation may include deleting metadata that associates the storage device with the logical disk (e.g., deleting one or more entries from a mapping of each storage device to one or more logical disks). It is noted that such dissociation from a logical disk need not affect other logical disks that are associated with the storage device. For example, if a storage device is associated with more than one logical disks, the storage device can be dissociated from one logical disks and remain associated with the other logical disks (i.e., disabling is not universal in the data center and is only on a per-logical disk or per-tenant basis). In other examples, the disassociating only by prevents a storage device from being a primary storage device in the logical disk and is only allowed to be a replica storage device in the logical disk. In this way, any “slower” storage device that has a negative effect on performance of the logical disk are only involved in write operations and are not involved in read operations, which can improve the overall performance of the logical disk.
At 614, adding a new storage device to the plurality of storage devices of the logical disk. The new storage device may a new instance of storage device that is added to the logical disk to reduce the load on others of the plurality of storage devices. In some examples, the new storage device is a storage disk that was previously removed (e.g., due to an operational failure) and is, again, added to the logical disk when the operational failure is resolved. As discussed above, in some examples, the addition of the storage device may be simulated to assess whether the addition will likely improve the performance of the storage disk.
At 618, pseudo-randomly redistribute objects across the plurality of storage devices. Such redistributing may be implemented using an algorithm to redistribute the objects. The redistribution occurs both when a storage device is added (to relocate some of the objects to the new storage device) and when a storage device is removed from the logical disk (to remove the objects from the storage device). In embodiments where simulation is used, the locations determined during the simulation may be used to relocate the objects (i.e., the algorithm is not executed again and, instead, the locations from the simulation are used). This has a benefit of causing the actual locations of the objects to match those determined in the simulation and, therefore, increases the likelihood of actual performance parameters matching the simulated performance parameters.
At 620, may loop from 618 back to calculating impact factors (at 606), which may be iterated for each of the plurality of storage disks comprising the logical disk. The logic 600 ends at 622. 622 may coincide with a start or end point of other logic, routines, and/or applications.
At a high level, the logic 600 may be used to reducing latency of a logical disk. Logic 600 may be implemented in a network element 126 (of
It is noted that any device (e.g., endpoint, network element, and the like) may execute logic 500 and/or 600. In some examples, the logic is implemented in instructions in a module (e.g., a logical disk management module) that has administrative rights to reconfigure logical disks. Such a module can be in a central controller, server, client module, and/or distributed to many network elements. For example, a central server (e.g., a network controller), coupled to the logical disk and the storage devices, may execute the logic. In other examples, the logic is distributed in small agents (e.g., administrative client) in servers that are coupled to the logical disk and the storage devices.
Turning to
To object F1.1, the server 704 assigns SD1 as the primary storage device and assigns SD2 and SD3 as the replica storage devices. To object F1.2, the server 704 assigns SD2 as the primary storage device and assigns SD1 and SD3 as the replica storage devices. To object F1.3, the server 704 assigns SD3 as its primary storage device and assigns SD1 and SD2 as the replica storage devices. To object F1.4, the server 704 assigns SD1 as the primary storage device and assigns SD2 and SD3 as the replica storage devices. To object F1.5, the server 704 assigns SD2 as the primary storage device and assigns SD1 and SD3 as its replica storage devices. At 720, the server 704 stores the objects F1.1 and F1.4 on SD1. At 724, SD1 transmits copies of the objects F1.1 and F1.4 to SD2 (e.g., based on SD2 being a replica storage device for the objects F1.1 and F1.4). At 726, SD1 transmits copies of the objects F1.1 and F1.4 to SD3 (e.g., based on SD3 being a replica storage device for the objects F1.1 and F1.4). At 728, the server 704 stores the objects F1.2 and F1.5 on the SD2. At 730, SD2 transmits copies of the objects F1.2 and F1.5 to SD1 (e.g., based on SD1 being a replica storage device for the objects F1.2 and F1.5). At 732, SD2 transmits copies of the objects F1.2 and F1.5 to SD3 (e.g., based on SD3 being a replica storage device for the objects F1.2 and F1.5). At 734, the server 704 stores the object F1.3 on the SD3. At 736, SD3 transmits copies of the object F1.3 to SD2 (e.g., based on SD2 being a replica storage device for the object F1.3). At 738, SD3 transmits copies of the object F1.3 to SD1 (e.g., based on SD1 being a replica storage device for the object F1.3). In this example, each primary storage device (e.g., using a distributed storage code block) copies an object to the appropriate replica storage device upon receipt of the object. However, in other examples, the server 704 may perform such distribution to replica storage devices while distributing objects to the primary storage devices. At 740, the server 704 transmits to the endpoint 702 (via the client) an acknowledgment that the File 1 was stored.
The latency X1 (as generally indicated by 722) is the time period between the endpoint 702 transmitting the File 1 for storage in the logical disk 705 and the File 1 being stored in the logical disk 705 (i.e., as objects F1.1-F1.5). The latency may be measured between the endpoint 702 transmitting the File 1 for storage in the logical disk 705 and acknowledgement (at 740) or at the completion of the storage of the last object (e.g., at 734, in this case). Latency X1 is an example of the latency of a write operation for the logical disk 705. The Latency X1 is influenced, at least in part, by a number of operations. Storing the file in the logical disk 705 causes each of the storage devices to execute multiple operations. In this example, at least nine operations are required to store the file in the logical disk 705 (i.e., 720, 724, 726, 728, 730, 732, 734, 736, and 738). Retrieving the file from the logical disk 705 causes each of the storage devices to execute one operation, as is illustrated in
Turning to
At 746, the server 704 determines objects comprising the file 1 and determines the storage location of each of the objects. In his example, the server determines that the File 1 was striped into five objects (i.e., F1.1-F1.5). The server 704 determines that object F1.1 is assigned SD1 as its primary storage device is assigned SD2 and SD3 as its replica storage devices. The server 704 determines that object F1.2 is assigned SD2 as its primary storage device is assigned SD1 and SD3 as its replica storage devices. The server 704 determines that object F1.3 is assigned SD3 as its primary storage device is assigned SD1 and SD2 as its replica storage devices. The server 704 determines that object F1.4 is assigned SD1 as its primary storage device is assigned SD2 and SD3 as its replica storage devices. The server 704 determines that object F1.5 is assigned SD2 as its primary storage device is assigned SD1 and SD3 as its replica storage devices.
The server 704 utilizes the logical disk 705 to retrieve each of the objects of the File 1 from their respective primary storage devices and not the replica storage devices. At 748, the server 704 retrieves the objects F1.1 and F1.4 from SD1 in the logical disk 705 (i.e., SD1 is the primary storage device for the objects F1.1 and F1.4). At 750, the server 704 retrieves the objects F1.2 and F1.5 from SD2 in the logical disk 705 (i.e., SD2 is the primary storage device for the objects F1.2 and F1.5). At 752, the server 704 retrieves the object F1.3 from SD3 in the logical disk 705 (i.e., SD3 is the primary storage device for the object F1.3).
At 754, the server 704 combines the objects F1.1, F1.2, F1.3, F1.4, and F1.5 into the File 1 (i.e., generates an instance of the File 1 from the objects). At 756, the server 704 transmits (via the client) the File 1 to the server endpoint 702.
The latency X2 (as generally indicated by 760) is the time period between the endpoint 702 requesting the File 1 from the logical disk 705 and the File 1 being transmitted to the endpoint 702. Latency X2 is an example of the latency of a read operation for the logical disk 705. The Latency X1 is influenced, at least in part, by a number of operations performed by the storage devices during the retrieval of the file. Retrieving the file from the logical disk 705 causes each of the storage devices to one operation. In this example, at least three operations are required to retrieve the file from the logical disk 705 (i.e., 748, 750, and 752).
Turning to
In some cases, reconfiguring the logical device 705 (e.g., by removing storage devices) may lead to others of the storage devices become overloaded with objects. Thus, a user interface (e.g., a graphical user interface (GUI), or command-line interface (CLI)) can be used to receive input from an endpoint associated with a user. The user interface allows a user to assess the impact of such reconfigurations and to approve or not approve and reconfigurations suggested by the server. In some embodiments, the threshold value is dynamically relaxed (e.g., becomes less restrictive) by a pre-specified amount (or percent) after each device is removed from the logical disk. Such dynamically relaxed threshold values helps reduce the likelihood of the degeneration of the logic disk due to the secondary impact of removing storage devices (repeatedly turning off a storage device, which results in others of the storage devices failing to meeting the threshold because they are sharing a higher proportion of the load than before the removal). For example, the threshold for latency may begin at 15 ms for any storage device in a logical disk. After one device is removed from the logical disk, the threshold is relaxed by a factor (e.g., 10%) and, therefore, becomes 15 ms*(1+0.1)=16.5 ms. After a second device is removed from the logical disk, the threshold is relaxed by the factor (e.g., 10%) and, therefore, becomes 16.5 ms*(1+0.1)=18.15 ms. In other cases the threshold is relaxed by an increment (e.g., 2 ms) and, therefore, can go from 15 ms to 15−2=13 ms (i.e., after one device is removed from the logical disk) and from 13 ms to 13−2=11 ms (i.e., after a second device is removed from the logical disk). In further embodiments, the threshold value is dynamically restricted (e.g., becomes more restrictive) by a pre-specified amount (or percent) after each device is removed from the logical disk.
Turning to
At 786, the server 704 determines an influence of each of the storage devices SD1, SD2, and SD3 on an average latency of the logical disk based on the performance parameters. For example, a numerical representation of the influence of each storage device may be determined using a mathematical model. In some examples, the numerical representation is a weighting factor for each of the plurality of storage devices SD1, SD2, and SD3. A portion of a latency of a logical device that is attributable to each storage device (i.e., a weighted latency for each storage device) in the logical device is calculated based, at least in part, on a corresponding latency of each storage device weighted by the corresponding weighting factor.
At 788, the server 704 calculates an average latency of the logical disk 705. A performance parameter of a logical disk can be calculated based, at least in part, on corresponding performance parameters each of the storage devices and the weighting factor. Thus, the latency of the logical disk 705 can be calculated by summing the weighted latency for each of the storage devices SD1, SD2, and SD3 in the logical disk 705.
At 790, the latency of the logical disk 705 and the influence of the storage devices SD1, SD2, and SD3 are transmitted to the endpoint 702. The latency of the logical disk 705 and the influence of the storage devices SD1, SD2, and SD3 may be rendered in the graphical interface (e.g., similar to that illustrate in
At 792, the endpoint 702 transmits to the server 704 a selection of one of the storage devices SD1, SD2, and SD3 to disable from the logical disk 705. The selection may be receives via the graphical interface. In this example, the selection identifies SD2 as the storage device to be removed from the logical disk 705. At 794, the server 704 simulates disabling the storage device from the logical disk 705. At 795, the server 704 transmits to the endpoint 702 a result of the simulation. The results may include simulated performance parameters of the logical disk 705 (e.g., with SD2 simulated as being removed). The graphical interface may generate a window for receiving, from the endpoint 702, input to confirm the original selection of SD2 at 792 (e.g., to accept or not accept the original selection based on the result of the simulation). At 796, the endpoint 702 transmits to the server 704 a confirmation of the selection (i.e., the original selection of SD2 at 792). At 798, the server 704 removes the storage device SD2 from the logical device 705. Removing SD2 improves the performance of a logical disk by reducing a latency of the logical disk 705. At 799, the server 704 pseudo-randomly redistributes objects from SD2 to SD1 and SD3.
Each of the graphical components (i.e., 802 and 804a-804f) is selectable to toggle on/or of a further display of detailed information associated with the storage device in the context of the logical disk. In this example, the detailed information includes a performance parameter and a performance impact for the selected storage device. The performance parameter is a latency (measured in milliseconds) of the selected storage device. The performance impact is a proportion of a latency of the logical device that is attributable to the selected storage device. In this example, each of the graphical components 804a and 804e were selected to display the further information in windows 806 and 814, respectively.
The window 806 includes text 810, text 812, and button 808. The text 810 identifies that the storage device OSD1 has a latency of 12 ms. The text 812 identifies that the storage device OSD1 has performance impact of 20 percent on the logical disk (i.e., the influence of the storage device on the logical disk). In other words, 20 percent of the latency of the logical device is attributable to the storage device OSD1. The button 808 includes the text “DISABLE”. When the button 808 is selected, it causes the corresponding storage device (in this case, OSD1) to be removed from the logical disk and causes the text to be togged from reading “DISABLE” to “ENABLE”. In effect, the button allows the corresponding storage device to be selectively removed from or added to the logical disk.
The window 814 includes text 818, text 820, and button 816. The text 818 identifies that the storage device OSD5 has a latency of 56 ms. It is noted, again, that the storage device OSD5 is not included in the logical disk (as indicated by the dashed line). The text 820 identifies a performance impact that resulted from simulating the storage device OSD5 being added to the logical disk. In this case, the storage device OSD5 would have a performance impact of 10 percent on the logical disk (i.e., the influence of the storage device on the logical disk). In other words, 10 percent of the latency of the logical device would be attributable to the storage device OSD5 (if it were added to the logical disk). The button 816 includes the text “ENABLE”. When the button 816 is selected, it causes the corresponding storage device (in this case, OSD5) to be added to the logical disk and causes the text to be togged from reading “ENABLE” to “DISABLE”.
The GUI 800 provides a device (e.g., an endpoint, network element, and the like) with interactive information describing the logical disk and the storage devices therein. For example, a user may use an input interface of the device (e.g., keyboard, a display, touchscreen, and/or of other input interface) to provide input to the GUI 800. Thus, the GUI 800 enables the device to control adding or removing storage devices from the logical disk.
In some examples, the graphical components 804a-804f may be rendered to graphical depict an indication of the influence of the storage device on the logical disk. For example, each of the graphical components 804a-804f may be shaded using a color that corresponds to their influence on the logical disk. In such an example, each of the graphical components 804a-804f is shaded with a color (e.g., filled with a color) selected from a gradient from a first color to a second color (e.g., where 0% influence corresponds to the first color and 100% influence corresponds to the second color). The gradient may be from white to black, green to red, or any other combination of first and second colors.
The example of
Additionally, it should be noted that with the examples provided above, interaction may be described in terms of specific numbers of (e.g., one, two, three, or four) network elements, endpoints, servers, logical disks, storage devices, etc. However, this has been done for purposes of clarity and example only. In certain cases, it may be easier to describe one or more of the functionalities of a given set of flows by only referencing a limited number of network elements, endpoints, servers, logical disks, storage devices, etc. It should be appreciated that the systems described herein are readily scalable and, further, can accommodate a large number of components, as well as more complicated/sophisticated arrangements and configurations. Accordingly, the examples provided should not limit the scope or inhibit the broad techniques of using various protocols for improving performance of object storage systems, as potentially applied to a myriad of other architectures.
It is also important to note that the steps in the Figures illustrate only some of the possible scenarios that may be executed by, or within, the elements described herein. Some of these steps may be deleted or removed where appropriate, or these steps may be modified or changed considerably without departing from the scope of the present disclosure. In addition, a number of these operations have been described as being executed concurrently with, or in parallel to, one or more additional operations. However, the timing of these operations may be altered considerably. The preceding operational flows have been offered for purposes of example and discussion. Substantial flexibility is provided by network elements, endpoints, servers, storage devices, in that any suitable arrangements, chronologies, configurations, and timing mechanisms may be provided without departing from the teachings of the present disclosure.
It should also be noted that many of the previous discussions may imply a single client-server relationship. In reality, there is a multitude of servers in the delivery tier in certain implementations of the present disclosure. Moreover, the present disclosure can readily be extended to apply to intervening servers further upstream in the architecture, though this is not necessarily correlated to the ‘m’ clients that are passing through the ‘n’ servers. Any such permutations, scaling, and configurations are clearly within the broad scope of the present disclosure.
Numerous other changes, substitutions, variations, alterations, and modifications may be ascertained to one skilled in the art and it is intended that the present disclosure encompass all such changes, substitutions, variations, alterations, and modifications as falling within the scope of the appended claims. In order to assist the United States Patent and Trademark Office (USPTO) and, additionally, any readers of any patent issued on this application in interpreting the claims appended hereto, Applicant wishes to note that the Applicant: (a) does not intend any of the appended claims to invoke paragraph six (6) of 35 U.S.C. section 112 as it exists on the date of the filing hereof unless the words “means for” or “step for” are specifically used in the particular claims; and (b) does not intend, by any statement in the specification, to limit this disclosure in any way that is not otherwise reflected in the appended claims.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4688695 | Hirohata | Aug 1987 | A |
| 5263003 | Cowles et al. | Nov 1993 | A |
| 5339445 | Gasztonyi | Aug 1994 | A |
| 5430859 | Norman et al. | Jul 1995 | A |
| 5457746 | Dolphin | Oct 1995 | A |
| 5535336 | Smith et al. | Jul 1996 | A |
| 5588012 | Oizumi | Dec 1996 | A |
| 5617421 | Chin et al. | Apr 1997 | A |
| 5680579 | Young et al. | Oct 1997 | A |
| 5690194 | Parker et al. | Nov 1997 | A |
| 5740171 | Mazzola et al. | Apr 1998 | A |
| 5742604 | Edsall et al. | Apr 1998 | A |
| 5764636 | Edsall | Jun 1998 | A |
| 5809285 | Hilland | Sep 1998 | A |
| 5812814 | Sukegawa | Sep 1998 | A |
| 5812950 | Tom | Sep 1998 | A |
| 5838970 | Thomas | Nov 1998 | A |
| 5999930 | Wolff | Dec 1999 | A |
| 6035105 | McCloghrie et al. | Mar 2000 | A |
| 6101497 | Ofek | Aug 2000 | A |
| 6148414 | Brown et al. | Nov 2000 | A |
| 6185203 | Berman | Feb 2001 | B1 |
| 6188694 | Fine et al. | Feb 2001 | B1 |
| 6202135 | Kedem et al. | Mar 2001 | B1 |
| 6208649 | Kloth | Mar 2001 | B1 |
| 6209059 | Ofer et al. | Mar 2001 | B1 |
| 6219699 | McCloghrie et al. | Apr 2001 | B1 |
| 6219753 | Richardson | Apr 2001 | B1 |
| 6223250 | Yokono | Apr 2001 | B1 |
| 6226771 | Hilla et al. | May 2001 | B1 |
| 6260120 | Blumenau et al. | Jul 2001 | B1 |
| 6266705 | Ullum et al. | Jul 2001 | B1 |
| 6269381 | St. Pierre et al. | Jul 2001 | B1 |
| 6269431 | Dunham | Jul 2001 | B1 |
| 6295575 | Blumenau et al. | Sep 2001 | B1 |
| 6400730 | Latif et al. | Jun 2002 | B1 |
| 6408406 | Parris | Jun 2002 | B1 |
| 6542909 | Tamer et al. | Apr 2003 | B1 |
| 6542961 | Matsunami et al. | Apr 2003 | B1 |
| 6553390 | Gross et al. | Apr 2003 | B1 |
| 6564252 | Hickman et al. | May 2003 | B1 |
| 6578121 | Schutzman | Jun 2003 | B1 |
| 6647474 | Yanai et al. | Nov 2003 | B2 |
| 6675258 | Bramhall et al. | Jan 2004 | B1 |
| 6683883 | Czeiger et al. | Jan 2004 | B1 |
| 6694413 | Mimatsu et al. | Feb 2004 | B1 |
| 6708227 | Cabrera et al. | Mar 2004 | B1 |
| 6711649 | Bachmat | Mar 2004 | B1 |
| 6715007 | Williams et al. | Mar 2004 | B1 |
| 6728791 | Young | Apr 2004 | B1 |
| 6772231 | Reuter et al. | Aug 2004 | B2 |
| 6820099 | Huber et al. | Nov 2004 | B1 |
| 6847647 | Wrenn | Jan 2005 | B1 |
| 6848759 | Doornbos et al. | Feb 2005 | B2 |
| 6850955 | Sonoda et al. | Feb 2005 | B2 |
| 6876656 | Brewer et al. | Apr 2005 | B2 |
| 6880062 | Ibrahim et al. | Apr 2005 | B1 |
| 6898670 | Nahum | May 2005 | B2 |
| 6907419 | Pesola et al. | Jun 2005 | B1 |
| 6912668 | Brown et al. | Jun 2005 | B1 |
| 6952734 | Gunlock et al. | Oct 2005 | B1 |
| 6976090 | Ben-Shaul et al. | Dec 2005 | B2 |
| 6978300 | Beukema et al. | Dec 2005 | B1 |
| 6983303 | Pellegrino et al. | Jan 2006 | B2 |
| 6986015 | Testardi | Jan 2006 | B2 |
| 6986069 | Oehler et al. | Jan 2006 | B2 |
| 7051056 | Rodriguez-Rivera et al. | May 2006 | B2 |
| 7069465 | Chu et al. | Jun 2006 | B2 |
| 7073017 | Yamamoto | Jul 2006 | B2 |
| 7108339 | Berger | Sep 2006 | B2 |
| 7149858 | Kiselev | Dec 2006 | B1 |
| 7171514 | Coronado et al. | Jan 2007 | B2 |
| 7171668 | Molloy et al. | Jan 2007 | B2 |
| 7174354 | Andreasson | Feb 2007 | B2 |
| 7200144 | Terrell et al. | Apr 2007 | B2 |
| 7222255 | Claessens et al. | May 2007 | B1 |
| 7237045 | Beckmann et al. | Jun 2007 | B2 |
| 7240188 | Takata et al. | Jul 2007 | B2 |
| 7246260 | Brown et al. | Jul 2007 | B2 |
| 7266718 | Idei et al. | Sep 2007 | B2 |
| 7269168 | Roy et al. | Sep 2007 | B2 |
| 7277431 | Walter et al. | Oct 2007 | B2 |
| 7277948 | Igarashi et al. | Oct 2007 | B2 |
| 7305658 | Hamilton et al. | Dec 2007 | B1 |
| 7328434 | Swanson et al. | Feb 2008 | B2 |
| 7340555 | Ashmore et al. | Mar 2008 | B2 |
| 7346751 | Prahlad et al. | Mar 2008 | B2 |
| 7352706 | Klotz et al. | Apr 2008 | B2 |
| 7353305 | Pangal et al. | Apr 2008 | B2 |
| 7359321 | Sindhu et al. | Apr 2008 | B1 |
| 7383381 | Faulkner et al. | Jun 2008 | B1 |
| 7433326 | Desai et al. | Oct 2008 | B2 |
| 7433948 | Edsall | Oct 2008 | B2 |
| 7434105 | Rodriguez-Rivera et al. | Oct 2008 | B1 |
| 7441154 | Klotz et al. | Oct 2008 | B2 |
| 7447839 | Uppala | Nov 2008 | B2 |
| 7487321 | Muthiah et al. | Feb 2009 | B2 |
| 7500053 | Kavuri et al. | Mar 2009 | B1 |
| 7542681 | Cornell et al. | Jun 2009 | B2 |
| 7558872 | Senevirathne et al. | Jul 2009 | B1 |
| 7587570 | Sarkar et al. | Sep 2009 | B2 |
| 7643505 | Colloff | Jan 2010 | B1 |
| 7654625 | Amann et al. | Feb 2010 | B2 |
| 7657796 | Kaiser et al. | Feb 2010 | B1 |
| 7668981 | Nagineni et al. | Feb 2010 | B1 |
| 7669071 | Cochran et al. | Feb 2010 | B2 |
| 7689384 | Becker | Mar 2010 | B1 |
| 7694092 | Mizuno | Apr 2010 | B2 |
| 7697554 | Ofer et al. | Apr 2010 | B1 |
| 7706303 | Bose et al. | Apr 2010 | B2 |
| 7707481 | Kirschner et al. | Apr 2010 | B2 |
| 7716648 | Vaidyanathan et al. | May 2010 | B2 |
| 7752360 | Galles | Jul 2010 | B2 |
| 7757059 | Ofer et al. | Jul 2010 | B1 |
| 7774329 | Peddy et al. | Aug 2010 | B1 |
| 7774839 | Nazzal | Aug 2010 | B2 |
| 7793138 | Rastogi et al. | Sep 2010 | B2 |
| 7840730 | D'Amato et al. | Nov 2010 | B2 |
| 7843906 | Chidambaram et al. | Nov 2010 | B1 |
| 7895428 | Boland, IV et al. | Feb 2011 | B2 |
| 7904599 | Bennett | Mar 2011 | B1 |
| 7930494 | Goheer et al. | Apr 2011 | B1 |
| 7975175 | Votta et al. | Jul 2011 | B2 |
| 7979670 | Saliba et al. | Jul 2011 | B2 |
| 7984259 | English | Jul 2011 | B1 |
| 8031703 | Gottumukkula et al. | Oct 2011 | B2 |
| 8032621 | Upalekar et al. | Oct 2011 | B1 |
| 8051197 | Mullendore et al. | Nov 2011 | B2 |
| 8086755 | Duffy, IV et al. | Dec 2011 | B2 |
| 8161134 | Mishra et al. | Apr 2012 | B2 |
| 8196018 | Forhan et al. | Jun 2012 | B2 |
| 8205951 | Boks | Jun 2012 | B2 |
| 8218538 | Chidambaram et al. | Jul 2012 | B1 |
| 8230066 | Heil | Jul 2012 | B2 |
| 8234377 | Cohn | Jul 2012 | B2 |
| 8266238 | Zimmer et al. | Sep 2012 | B2 |
| 8272104 | Chen et al. | Sep 2012 | B2 |
| 8274993 | Sharma et al. | Sep 2012 | B2 |
| 8290919 | Kelly et al. | Oct 2012 | B1 |
| 8297722 | Chambers et al. | Oct 2012 | B2 |
| 8301746 | Head et al. | Oct 2012 | B2 |
| 8335231 | Kloth et al. | Dec 2012 | B2 |
| 8341121 | Claudatos et al. | Dec 2012 | B1 |
| 8345692 | Smith | Jan 2013 | B2 |
| 8352941 | Protopopov et al. | Jan 2013 | B1 |
| 8392760 | Kandula et al. | Mar 2013 | B2 |
| 8442059 | de la Iglesia et al. | May 2013 | B1 |
| 8479211 | Marshall et al. | Jul 2013 | B1 |
| 8495356 | Ashok et al. | Jul 2013 | B2 |
| 8514868 | Hill | Aug 2013 | B2 |
| 8532108 | Li et al. | Sep 2013 | B2 |
| 8560663 | Baucke et al. | Oct 2013 | B2 |
| 8619599 | Even | Dec 2013 | B1 |
| 8626891 | Guru et al. | Jan 2014 | B2 |
| 8630983 | Sengupta et al. | Jan 2014 | B2 |
| 8660129 | Brendel et al. | Feb 2014 | B1 |
| 8661299 | Ip | Feb 2014 | B1 |
| 8677485 | Sharma et al. | Mar 2014 | B2 |
| 8683296 | Anderson et al. | Mar 2014 | B2 |
| 8706772 | Hartig et al. | Apr 2014 | B2 |
| 8719804 | Jain | May 2014 | B2 |
| 8725854 | Edsall | May 2014 | B2 |
| 8768981 | Milne et al. | Jul 2014 | B1 |
| 8775773 | Acharya et al. | Jul 2014 | B2 |
| 8793372 | Ashok et al. | Jul 2014 | B2 |
| 8805918 | Chandrasekaran et al. | Aug 2014 | B1 |
| 8805951 | Faibish et al. | Aug 2014 | B1 |
| 8832330 | Lancaster | Sep 2014 | B1 |
| 8855116 | Rosset et al. | Oct 2014 | B2 |
| 8856339 | Mestery et al. | Oct 2014 | B2 |
| 8868474 | Leung et al. | Oct 2014 | B2 |
| 8887286 | Dupont et al. | Nov 2014 | B2 |
| 8898385 | Jayaraman et al. | Nov 2014 | B2 |
| 8909928 | Ahmad et al. | Dec 2014 | B2 |
| 8918510 | Gmach et al. | Dec 2014 | B2 |
| 8924720 | Raghuram et al. | Dec 2014 | B2 |
| 8930747 | Levijarvi et al. | Jan 2015 | B2 |
| 8935500 | Gulati et al. | Jan 2015 | B1 |
| 8949677 | Brundage et al. | Feb 2015 | B1 |
| 8996837 | Bono et al. | Mar 2015 | B1 |
| 9003086 | Schuller et al. | Apr 2015 | B1 |
| 9007922 | Mittal et al. | Apr 2015 | B1 |
| 9009427 | Sharma et al. | Apr 2015 | B2 |
| 9009704 | McGrath et al. | Apr 2015 | B2 |
| 9026492 | Shorey | May 2015 | B1 |
| 9075638 | Barnett et al. | Jul 2015 | B2 |
| 9141554 | Candelaria | Sep 2015 | B1 |
| 9141785 | Mukkara et al. | Sep 2015 | B2 |
| 9164795 | Vincent | Oct 2015 | B1 |
| 9176677 | Fradkin et al. | Nov 2015 | B1 |
| 9201704 | Chang et al. | Dec 2015 | B2 |
| 9203784 | Chang et al. | Dec 2015 | B2 |
| 9207882 | Rosset et al. | Dec 2015 | B2 |
| 9207929 | Katsura | Dec 2015 | B2 |
| 9213612 | Candelaria | Dec 2015 | B2 |
| 9223564 | Munireddy et al. | Dec 2015 | B2 |
| 9223634 | Chang et al. | Dec 2015 | B2 |
| 9244761 | Yekhanin et al. | Jan 2016 | B2 |
| 9250969 | Lager-Cavilla et al. | Feb 2016 | B2 |
| 9264494 | Factor et al. | Feb 2016 | B2 |
| 9270754 | Iyengar et al. | Feb 2016 | B2 |
| 9280487 | Candelaria | Mar 2016 | B2 |
| 9304815 | Vasanth et al. | Apr 2016 | B1 |
| 9313048 | Chang et al. | Apr 2016 | B2 |
| 9374270 | Nakil et al. | Jun 2016 | B2 |
| 9378060 | Jansson et al. | Jun 2016 | B2 |
| 9396251 | Boudreau et al. | Jul 2016 | B1 |
| 9448877 | Candelaria | Sep 2016 | B2 |
| 9471348 | Zuo et al. | Oct 2016 | B2 |
| 9501473 | Kong et al. | Nov 2016 | B1 |
| 9503523 | Rosset et al. | Nov 2016 | B2 |
| 9565110 | Mullendore et al. | Feb 2017 | B2 |
| 9575828 | Agarwal et al. | Feb 2017 | B2 |
| 9614763 | Dong et al. | Apr 2017 | B2 |
| 9658868 | Hill | May 2017 | B2 |
| 9658876 | Chang et al. | May 2017 | B2 |
| 9733868 | Chandrasekaran et al. | Aug 2017 | B2 |
| 9763518 | Charest et al. | Sep 2017 | B2 |
| 9830240 | George et al. | Nov 2017 | B2 |
| 9853873 | Dasu et al. | Dec 2017 | B2 |
| 20020049980 | Hoang | Apr 2002 | A1 |
| 20020053009 | Selkirk et al. | May 2002 | A1 |
| 20020073276 | Howard et al. | Jun 2002 | A1 |
| 20020083120 | Soltis | Jun 2002 | A1 |
| 20020095547 | Watanabe et al. | Jul 2002 | A1 |
| 20020103889 | Markson et al. | Aug 2002 | A1 |
| 20020103943 | Lo et al. | Aug 2002 | A1 |
| 20020112113 | Karpoff et al. | Aug 2002 | A1 |
| 20020120741 | Webb et al. | Aug 2002 | A1 |
| 20020138675 | Mann | Sep 2002 | A1 |
| 20020156971 | Jones | Oct 2002 | A1 |
| 20030023885 | Potter et al. | Jan 2003 | A1 |
| 20030026267 | Oberman et al. | Feb 2003 | A1 |
| 20030055933 | Ishizaki et al. | Mar 2003 | A1 |
| 20030056126 | O'Connor et al. | Mar 2003 | A1 |
| 20030065986 | Fraenkel et al. | Apr 2003 | A1 |
| 20030084359 | Bresniker et al. | May 2003 | A1 |
| 20030118053 | Edsall et al. | Jun 2003 | A1 |
| 20030131105 | Czeiger et al. | Jul 2003 | A1 |
| 20030131165 | Asano et al. | Jul 2003 | A1 |
| 20030131182 | Kumar et al. | Jul 2003 | A1 |
| 20030140134 | Swanson et al. | Jul 2003 | A1 |
| 20030140210 | Testardi | Jul 2003 | A1 |
| 20030149763 | Heitman et al. | Aug 2003 | A1 |
| 20030154271 | Baldwin et al. | Aug 2003 | A1 |
| 20030159058 | Eguchi et al. | Aug 2003 | A1 |
| 20030174725 | Shankar | Sep 2003 | A1 |
| 20030189395 | Doornbos et al. | Oct 2003 | A1 |
| 20030210686 | Terrell et al. | Nov 2003 | A1 |
| 20040024961 | Cochran et al. | Feb 2004 | A1 |
| 20040030857 | Krakirian et al. | Feb 2004 | A1 |
| 20040039939 | Cox et al. | Feb 2004 | A1 |
| 20040054776 | Klotz et al. | Mar 2004 | A1 |
| 20040057389 | Klotz et al. | Mar 2004 | A1 |
| 20040059807 | Klotz et al. | Mar 2004 | A1 |
| 20040088574 | Walter et al. | May 2004 | A1 |
| 20040117438 | Considine et al. | Jun 2004 | A1 |
| 20040123029 | Dalai et al. | Jun 2004 | A1 |
| 20040128470 | Hetzler et al. | Jul 2004 | A1 |
| 20040153863 | Klotz et al. | Aug 2004 | A1 |
| 20040215749 | Tsao | Oct 2004 | A1 |
| 20040230848 | Mayo et al. | Nov 2004 | A1 |
| 20040250034 | Yagawa et al. | Dec 2004 | A1 |
| 20050033936 | Nakano et al. | Feb 2005 | A1 |
| 20050036499 | Dutt et al. | Feb 2005 | A1 |
| 20050050211 | Kaul et al. | Mar 2005 | A1 |
| 20050050270 | Horn | Mar 2005 | A1 |
| 20050053073 | Kloth et al. | Mar 2005 | A1 |
| 20050055428 | Terai et al. | Mar 2005 | A1 |
| 20050060574 | Klotz et al. | Mar 2005 | A1 |
| 20050060598 | Klotz et al. | Mar 2005 | A1 |
| 20050071851 | Opheim | Mar 2005 | A1 |
| 20050076113 | Klotz et al. | Apr 2005 | A1 |
| 20050091426 | Horn et al. | Apr 2005 | A1 |
| 20050114615 | Ogasawara et al. | May 2005 | A1 |
| 20050117522 | Basavaiah et al. | Jun 2005 | A1 |
| 20050117562 | Wrenn | Jun 2005 | A1 |
| 20050138287 | Ogasawara et al. | Jun 2005 | A1 |
| 20050185597 | Le et al. | Aug 2005 | A1 |
| 20050188170 | Yamamoto | Aug 2005 | A1 |
| 20050235072 | Smith et al. | Oct 2005 | A1 |
| 20050283658 | Clark et al. | Dec 2005 | A1 |
| 20060015861 | Takata et al. | Jan 2006 | A1 |
| 20060015928 | Setty et al. | Jan 2006 | A1 |
| 20060034302 | Peterson | Feb 2006 | A1 |
| 20060045021 | Deragon et al. | Mar 2006 | A1 |
| 20060098672 | Schzukin et al. | May 2006 | A1 |
| 20060117099 | Mogul | Jun 2006 | A1 |
| 20060136684 | Le et al. | Jun 2006 | A1 |
| 20060184287 | Belady et al. | Aug 2006 | A1 |
| 20060198319 | Schondelmayer et al. | Sep 2006 | A1 |
| 20060215297 | Kikuchi | Sep 2006 | A1 |
| 20060230227 | Ogasawara et al. | Oct 2006 | A1 |
| 20060242332 | Johnsen et al. | Oct 2006 | A1 |
| 20060251111 | Kloth et al. | Nov 2006 | A1 |
| 20070005297 | Beresniewicz et al. | Jan 2007 | A1 |
| 20070067593 | Satoyama et al. | Mar 2007 | A1 |
| 20070079068 | Draggon | Apr 2007 | A1 |
| 20070094465 | Sharma et al. | Apr 2007 | A1 |
| 20070101202 | Garbow | May 2007 | A1 |
| 20070121519 | Cuni et al. | May 2007 | A1 |
| 20070136541 | Herz et al. | Jun 2007 | A1 |
| 20070162969 | Becker | Jul 2007 | A1 |
| 20070211640 | Palacharla et al. | Sep 2007 | A1 |
| 20070214316 | Kim | Sep 2007 | A1 |
| 20070250838 | Belady et al. | Oct 2007 | A1 |
| 20070263545 | Foster et al. | Nov 2007 | A1 |
| 20070276884 | Hara et al. | Nov 2007 | A1 |
| 20070283059 | Ho et al. | Dec 2007 | A1 |
| 20080016412 | White et al. | Jan 2008 | A1 |
| 20080034149 | Sheen | Feb 2008 | A1 |
| 20080052459 | Chang et al. | Feb 2008 | A1 |
| 20080059698 | Kabir et al. | Mar 2008 | A1 |
| 20080114933 | Ogasawara et al. | May 2008 | A1 |
| 20080126509 | Subrannanian et al. | May 2008 | A1 |
| 20080126734 | Murase | May 2008 | A1 |
| 20080168304 | Flynn et al. | Jul 2008 | A1 |
| 20080201616 | Ashmore | Aug 2008 | A1 |
| 20080209044 | Forrester | Aug 2008 | A1 |
| 20080244184 | Lewis et al. | Oct 2008 | A1 |
| 20080256082 | Davies et al. | Oct 2008 | A1 |
| 20080267217 | Colville et al. | Oct 2008 | A1 |
| 20080288661 | Galles | Nov 2008 | A1 |
| 20080294888 | Ando et al. | Nov 2008 | A1 |
| 20090063766 | Matsumura et al. | Mar 2009 | A1 |
| 20090083484 | Basham et al. | Mar 2009 | A1 |
| 20090089567 | Boland, IV et al. | Apr 2009 | A1 |
| 20090094380 | Qiu et al. | Apr 2009 | A1 |
| 20090094664 | Butler et al. | Apr 2009 | A1 |
| 20090125694 | Innan et al. | May 2009 | A1 |
| 20090193223 | Saliba et al. | Jul 2009 | A1 |
| 20090201926 | Kagan et al. | Aug 2009 | A1 |
| 20090222733 | Basham et al. | Sep 2009 | A1 |
| 20090240873 | Yu et al. | Sep 2009 | A1 |
| 20090282471 | Green et al. | Nov 2009 | A1 |
| 20090323706 | Germain et al. | Dec 2009 | A1 |
| 20100011365 | Gerovac et al. | Jan 2010 | A1 |
| 20100030995 | Wang et al. | Feb 2010 | A1 |
| 20100046378 | Knapp et al. | Feb 2010 | A1 |
| 20100083055 | Ozonat | Apr 2010 | A1 |
| 20100174968 | Charles et al. | Jul 2010 | A1 |
| 20100318609 | Lahiri et al. | Dec 2010 | A1 |
| 20100318837 | Murphy et al. | Dec 2010 | A1 |
| 20110010394 | Carew et al. | Jan 2011 | A1 |
| 20110022691 | Banerjee et al. | Jan 2011 | A1 |
| 20110029824 | Schöler et al. | Feb 2011 | A1 |
| 20110035494 | Pandey et al. | Feb 2011 | A1 |
| 20110075667 | Li et al. | Mar 2011 | A1 |
| 20110087848 | Trent | Apr 2011 | A1 |
| 20110119556 | de Buen | May 2011 | A1 |
| 20110142053 | Van Der Merwe et al. | Jun 2011 | A1 |
| 20110173303 | Rider | Jul 2011 | A1 |
| 20110228679 | Varma et al. | Sep 2011 | A1 |
| 20110231899 | Pulier et al. | Sep 2011 | A1 |
| 20110239039 | Dieffenbach et al. | Sep 2011 | A1 |
| 20110252274 | Kawaguchi et al. | Oct 2011 | A1 |
| 20110276584 | Cotner et al. | Nov 2011 | A1 |
| 20110276675 | Singh et al. | Nov 2011 | A1 |
| 20110276951 | Jain | Nov 2011 | A1 |
| 20110299539 | Rajagopal et al. | Dec 2011 | A1 |
| 20110307450 | Hahn et al. | Dec 2011 | A1 |
| 20110313973 | Srivas et al. | Dec 2011 | A1 |
| 20120023319 | Chin et al. | Jan 2012 | A1 |
| 20120030401 | Cowan et al. | Feb 2012 | A1 |
| 20120054367 | Ramakrishnan et al. | Mar 2012 | A1 |
| 20120072578 | Alam | Mar 2012 | A1 |
| 20120072985 | Davne et al. | Mar 2012 | A1 |
| 20120084445 | Brock et al. | Apr 2012 | A1 |
| 20120084782 | Chou et al. | Apr 2012 | A1 |
| 20120096134 | Suit | Apr 2012 | A1 |
| 20120130874 | Mane et al. | May 2012 | A1 |
| 20120131174 | Ferris et al. | May 2012 | A1 |
| 20120134672 | Banerjee | May 2012 | A1 |
| 20120144014 | Natham et al. | Jun 2012 | A1 |
| 20120167094 | Suit | Jun 2012 | A1 |
| 20120173581 | Hartig et al. | Jul 2012 | A1 |
| 20120173589 | Kwon et al. | Jul 2012 | A1 |
| 20120177039 | Berman | Jul 2012 | A1 |
| 20120177041 | Berman | Jul 2012 | A1 |
| 20120177042 | Berman | Jul 2012 | A1 |
| 20120177043 | Berman | Jul 2012 | A1 |
| 20120177044 | Berman | Jul 2012 | A1 |
| 20120177045 | Berman | Jul 2012 | A1 |
| 20120177370 | Berman | Jul 2012 | A1 |
| 20120179909 | Sagi et al. | Jul 2012 | A1 |
| 20120210041 | Flynn et al. | Aug 2012 | A1 |
| 20120254440 | Wang | Oct 2012 | A1 |
| 20120265976 | Spiers et al. | Oct 2012 | A1 |
| 20120281706 | Agarwal et al. | Nov 2012 | A1 |
| 20120297088 | Wang et al. | Nov 2012 | A1 |
| 20120303618 | Dutta et al. | Nov 2012 | A1 |
| 20120311106 | Morgan | Dec 2012 | A1 |
| 20120311568 | Jansen | Dec 2012 | A1 |
| 20120320788 | Venkataramanan et al. | Dec 2012 | A1 |
| 20120324114 | Dutta et al. | Dec 2012 | A1 |
| 20120331119 | Bose et al. | Dec 2012 | A1 |
| 20130003737 | Sinicrope | Jan 2013 | A1 |
| 20130013664 | Baird et al. | Jan 2013 | A1 |
| 20130028135 | Berman | Jan 2013 | A1 |
| 20130036212 | Jibbe et al. | Feb 2013 | A1 |
| 20130036213 | Hasan et al. | Feb 2013 | A1 |
| 20130036449 | Mukkara et al. | Feb 2013 | A1 |
| 20130054888 | Bhat et al. | Feb 2013 | A1 |
| 20130061089 | Valiyaparambil et al. | Mar 2013 | A1 |
| 20130067162 | Jayaraman et al. | Mar 2013 | A1 |
| 20130080823 | Roth et al. | Mar 2013 | A1 |
| 20130086340 | Fleming et al. | Apr 2013 | A1 |
| 20130100858 | Kamath et al. | Apr 2013 | A1 |
| 20130111540 | Sabin | May 2013 | A1 |
| 20130138816 | Kuo et al. | May 2013 | A1 |
| 20130138836 | Cohen et al. | May 2013 | A1 |
| 20130139138 | Kakos | May 2013 | A1 |
| 20130144933 | Hinni et al. | Jun 2013 | A1 |
| 20130152076 | Patel | Jun 2013 | A1 |
| 20130152175 | Hromoko et al. | Jun 2013 | A1 |
| 20130163426 | Beliveau et al. | Jun 2013 | A1 |
| 20130163606 | Bagepalli et al. | Jun 2013 | A1 |
| 20130179941 | McGloin et al. | Jul 2013 | A1 |
| 20130182712 | Aguayo et al. | Jul 2013 | A1 |
| 20130185433 | Zhu et al. | Jul 2013 | A1 |
| 20130191106 | Kephart et al. | Jul 2013 | A1 |
| 20130198730 | Munireddy et al. | Aug 2013 | A1 |
| 20130208888 | Agrawal et al. | Aug 2013 | A1 |
| 20130212130 | Rahnama | Aug 2013 | A1 |
| 20130238641 | Mandelstein et al. | Sep 2013 | A1 |
| 20130266307 | Garg et al. | Oct 2013 | A1 |
| 20130268922 | Tiwari et al. | Oct 2013 | A1 |
| 20130275470 | Cao et al. | Oct 2013 | A1 |
| 20130297655 | Narasayya et al. | Nov 2013 | A1 |
| 20130297769 | Chang et al. | Nov 2013 | A1 |
| 20130318134 | Bolik et al. | Nov 2013 | A1 |
| 20130318288 | Khan et al. | Nov 2013 | A1 |
| 20140006708 | Huynh et al. | Jan 2014 | A1 |
| 20140016493 | Johnsson et al. | Jan 2014 | A1 |
| 20140019684 | Wei | Jan 2014 | A1 |
| 20140025770 | Warfield et al. | Jan 2014 | A1 |
| 20140029441 | Nydell | Jan 2014 | A1 |
| 20140029442 | Wallman | Jan 2014 | A1 |
| 20140039683 | Zimmermann et al. | Feb 2014 | A1 |
| 20140040473 | Ho et al. | Feb 2014 | A1 |
| 20140040883 | Tompkins | Feb 2014 | A1 |
| 20140047201 | Mehta | Feb 2014 | A1 |
| 20140053264 | Dubrovsky et al. | Feb 2014 | A1 |
| 20140059187 | Rosset et al. | Feb 2014 | A1 |
| 20140059266 | Ben-Michael et al. | Feb 2014 | A1 |
| 20140086253 | Yong | Mar 2014 | A1 |
| 20140089273 | Borshack et al. | Mar 2014 | A1 |
| 20140096249 | Dupont et al. | Apr 2014 | A1 |
| 20140105009 | Vos et al. | Apr 2014 | A1 |
| 20140108474 | David et al. | Apr 2014 | A1 |
| 20140109071 | Ding et al. | Apr 2014 | A1 |
| 20140112122 | Kapadia et al. | Apr 2014 | A1 |
| 20140123207 | Agarwal et al. | May 2014 | A1 |
| 20140156557 | Zeng et al. | Jun 2014 | A1 |
| 20140164666 | Yand | Jun 2014 | A1 |
| 20140164866 | Bolotov et al. | Jun 2014 | A1 |
| 20140172371 | Zhu et al. | Jun 2014 | A1 |
| 20140173060 | Jubran et al. | Jun 2014 | A1 |
| 20140173195 | Rosset et al. | Jun 2014 | A1 |
| 20140173579 | McDonald et al. | Jun 2014 | A1 |
| 20140189278 | Peng | Jul 2014 | A1 |
| 20140198794 | Mehta et al. | Jul 2014 | A1 |
| 20140211661 | Gorkemli et al. | Jul 2014 | A1 |
| 20140215265 | Mohanta et al. | Jul 2014 | A1 |
| 20140215590 | Brand | Jul 2014 | A1 |
| 20140219086 | Cantu′ et al. | Aug 2014 | A1 |
| 20140222953 | Karve et al. | Aug 2014 | A1 |
| 20140229790 | Goss et al. | Aug 2014 | A1 |
| 20140244585 | Sivasubramanian et al. | Aug 2014 | A1 |
| 20140244897 | Goss et al. | Aug 2014 | A1 |
| 20140245435 | Belenky | Aug 2014 | A1 |
| 20140269390 | Ciodaru et al. | Sep 2014 | A1 |
| 20140281700 | Nagesharao et al. | Sep 2014 | A1 |
| 20140297941 | Rajani et al. | Oct 2014 | A1 |
| 20140307578 | DeSanti | Oct 2014 | A1 |
| 20140317206 | Lomelino et al. | Oct 2014 | A1 |
| 20140324862 | Bingham et al. | Oct 2014 | A1 |
| 20140325208 | Resch et al. | Oct 2014 | A1 |
| 20140331276 | Frascadore et al. | Nov 2014 | A1 |
| 20140348166 | Yang et al. | Nov 2014 | A1 |
| 20140355450 | Bhikkaji et al. | Dec 2014 | A1 |
| 20140366155 | Chang et al. | Dec 2014 | A1 |
| 20150003450 | Salam et al. | Jan 2015 | A1 |
| 20150003458 | Li et al. | Jan 2015 | A1 |
| 20150003463 | Li et al. | Jan 2015 | A1 |
| 20150010001 | Duda et al. | Jan 2015 | A1 |
| 20150016461 | Qiang | Jan 2015 | A1 |
| 20150030024 | Venkataswami et al. | Jan 2015 | A1 |
| 20150046123 | Kato | Feb 2015 | A1 |
| 20150063353 | Kapadia et al. | Mar 2015 | A1 |
| 20150067001 | Koltsidas | Mar 2015 | A1 |
| 20150082432 | Eaton et al. | Mar 2015 | A1 |
| 20150120907 | Niestemski et al. | Apr 2015 | A1 |
| 20150121131 | Kiselev et al. | Apr 2015 | A1 |
| 20150127979 | Doppalapudi | May 2015 | A1 |
| 20150169313 | Katsura | Jun 2015 | A1 |
| 20150180672 | Kuwata | Jun 2015 | A1 |
| 20150207763 | Bertran Ortiz et al. | Jun 2015 | A1 |
| 20150205974 | Talley et al. | Jul 2015 | A1 |
| 20150222444 | Sarkar | Aug 2015 | A1 |
| 20150229546 | Somaiya et al. | Aug 2015 | A1 |
| 20150248366 | Bergsten et al. | Sep 2015 | A1 |
| 20150248418 | Bhardwaj et al. | Sep 2015 | A1 |
| 20150254003 | Lee et al. | Sep 2015 | A1 |
| 20150254088 | Chou et al. | Sep 2015 | A1 |
| 20150261446 | Lee | Sep 2015 | A1 |
| 20150263993 | Kuch et al. | Sep 2015 | A1 |
| 20150269048 | Marr et al. | Sep 2015 | A1 |
| 20150281067 | Wu | Oct 2015 | A1 |
| 20150303949 | Jafarkhani et al. | Oct 2015 | A1 |
| 20150341237 | Cuni et al. | Nov 2015 | A1 |
| 20150341239 | Bertran Ortiz et al. | Nov 2015 | A1 |
| 20150358136 | Medard | Dec 2015 | A1 |
| 20150379150 | Duda | Dec 2015 | A1 |
| 20160011936 | Luby | Jan 2016 | A1 |
| 20160011942 | Golbourn et al. | Jan 2016 | A1 |
| 20160062820 | Jones et al. | Mar 2016 | A1 |
| 20160070652 | Sundararaman et al. | Mar 2016 | A1 |
| 20160087885 | Tripathi et al. | Mar 2016 | A1 |
| 20160088083 | Bharadwaj et al. | Mar 2016 | A1 |
| 20160119159 | Zhao et al. | Apr 2016 | A1 |
| 20160139820 | Fluman | May 2016 | A1 |
| 20160205189 | Mopur et al. | Jul 2016 | A1 |
| 20160210161 | Rosset et al. | Jul 2016 | A1 |
| 20160274926 | Narasimhamurthy et al. | Sep 2016 | A1 |
| 20160285760 | Dong | Sep 2016 | A1 |
| 20160292359 | Tellis et al. | Oct 2016 | A1 |
| 20160294983 | Kliteynik et al. | Oct 2016 | A1 |
| 20160334998 | George et al. | Nov 2016 | A1 |
| 20160366094 | Mason et al. | Dec 2016 | A1 |
| 20160378624 | Jenkins, Jr. et al. | Dec 2016 | A1 |
| 20160380694 | Guduru | Dec 2016 | A1 |
| 20170010874 | Rosset | Jan 2017 | A1 |
| 20170010930 | Dutta et al. | Jan 2017 | A1 |
| 20170019475 | Metz et al. | Jan 2017 | A1 |
| 20170068630 | Iskandar et al. | Mar 2017 | A1 |
| 20170168970 | Sajeepa et al. | Jun 2017 | A1 |
| 20170177860 | Suarez et al. | Jun 2017 | A1 |
| 20170212858 | Chu et al. | Jul 2017 | A1 |
| 20170277655 | Das et al. | Sep 2017 | A1 |
| 20170337097 | Sipos et al. | Nov 2017 | A1 |
| 20170340113 | Charest et al. | Nov 2017 | A1 |
| Number | Date | Country |
|---|---|---|
| 2228719 | Sep 2010 | EP |
| 2439637 | Apr 2012 | EP |
| 2680155 | Jan 2014 | EP |
| 2350028 | May 2001 | GB |
| 2000-242434 | Sep 2000 | JP |
| 1566104 | Jan 2017 | TW |
| WO 2004077214 | Sep 2004 | WO |
| WO 2016003408 | Jan 2016 | WO |
| WO 2016003489 | Jan 2016 | WO |
| Entry |
|---|
| Xue, C, et al. “A Standard framework for Ceph performance profiling with latency breakdown,” CEPH, Jun. 30, 2015, 3 pages; https://wiki.ceph.com/Planning/Blueprints/Infernalis/A_standard_framework_for_Ceph_performance_profiling_with_latency_break. |
| Weil, S., et al. “Ceph: A scalable, high-performance distributed file system,” Proceedings of the 7th symposium on Operating systems design and implementation. USENIX Association, Nov. 6, 2006, 14 pages; http://www.ssrc.ucsc.edu/Papers/weil-osdi06.pdf. |
| Weil, Sage A., et al. “CRUSH: Controlled, scalable, decentralized placement of replicated data.” Proceedings of the 2006 ACM/IEEE conference on Supercomputing. ACM, Nov. 11, 2006, 12 pages; http://sc06.supercomp.org/schedule/pdf/pap107.pdf. |
| Hatzieleftheriou, A., et al., “Host-side Filesystem Journaling for Durable Shared Storage,” 13th USENIX Conference on File and Storage Technologies (FAST '15), Feb. 16-19, 2015, 9 pages; https://www.usenix.org/system/files/conference/fast15/fast15-paper-hatzieleftheriou.pdf. |
| Wang, F., et al. “OBFS: A File System for Object-Based Storage Devices,” Storage System Research Center, MSST. vol. 4., Apr. 2004, 18 pages. |
| Wu, J., et al., “The Design, and Implementation of AQuA: An Adaptive Quality of Service Aware Object-Based Storage Device,” Department of Computer Science, MSST, May 17, 2006, 25 pages; http://storageconference.us/2006/Presentations/30Wu.pdf. |
| Author Unknown, “5 Benefits of a Storage Gateway in the Cloud,” Blog, TwinStrata, Inc., posted Jul. 10, 2012, 4 pages, https://web.archive.org/web/20120725092619/http://blog.twinstrata.com/2012/07/10//5-benefits-of-a-storage-gateway-in-the-cloud. |
| Author Unknown, “Configuration Interface for IBM System Storage DS5000, IBM DS4000, and IBM DS3000 Systems,” IBM SAN Volume Controller Version 7.1, IBM® System Storage® SAN Volume Controller Information Center, Jun. 16, 2013, 3 pages. |
| Author Unknown, “Coraid EtherCloud, Software-Defined Storage with Scale-Out Infrastructure,” Solution Brief, 2013, 2 pages, Coraid, Redwood City, California, U.S.A. |
| Author Unknown, “Coraid Virtual DAS (VDAS) Technology: Eliminate Tradeoffs between DAS and Networked Storage,” Coraid Technology Brief, © 2013 Cora id, Inc., Published on or about Mar. 20, 2013, 2 pages. |
| Author Unknown, “Creating Performance-based SAN SLAs Using Finisar's NetWisdom” May 2006, 7 pages, Finisar Corporation, Sunnyvale, California, U.S.A. |
| Author Unknown, “Data Center, Metro Cloud Connectivity: Integrated Metro SAN Connectivity in 16 Gbps Switches,” Brocade Communication Systems, Inc., Apr. 2011, 14 pages. |
| Author Unknown, “Data Center, SAN Fabric Administration Best Practices Guide, Support Perspective,” Brocade Communication Systems, Inc., May 2013, 21 pages. |
| Author Unknown, “delphi—Save a CRC value in a file, without altering the actual CRC Checksum?” Stack Overflow, stackoverflow.com, Dec. 23, 2011, XP055130879, 3 pages http://stackoverflow.com/questions/8608219/save-a-crc-value-in-a-file-without-altering-the-actual-crc-checksum. |
| Author Unknown, “EMC Unisphere: Innovative Approach to Managing Low-End and Midrange Storage; Redefining Simplicity in the Entry-Level and Midrange Storage Markets,” Data sheet, EMC Corporation; published on or about Jan. 4, 2013 [Retrieved and printed Sep. 12, 2013] 6 pages http://www.emc.com/storage/vnx/unisphere.htm. |
| Author Unknown, “HP XP Array Manager Software—Overview & Features,” Storage Device Management Software; Hewlett-Packard Development Company, 3 pages; © 2013 Hewlett-Packard Development Company, L.P. |
| Author Unknown, “Joint Cisco and VMWare Solution for Optimizing Virtual Desktop Delivery: Data Center 3.0: Solutions to Accelerate Data Center Virtualization,” Cisco Systems, Inc. and VMware, Inc., Sep. 2008, 10 pages. |
| Author Unknown, “Network Transformation with Software-Defined Networking and Ethernet Fabrics,” Positioning Paper, 2012, 6 pages, Brocade Communications Systems. |
| Author Unknown, “Recreating Real Application Traffic in Junosphere Lab,” Solution Brief, Juniper Networks, Dec. 2011, 3 pages. |
| Author Unknown, “Shunra for HP Softwarer,” Enabiling Confidence in Application Performance Before Deployment, 2010, 2 pages. |
| Author Unknown, “Software Defined Networking: The New Norm for Networks,” White Paper, Open Networking Foundation, Apr. 13, 2012, 12 pages. |
| Author Unknown, “Software Defined Storage Networks an Introduction,” White Paper, Doc # 01-000030-001 Rev. A, Dec. 12, 2012, 8 pages; Jeda Networks, Newport Beach, California, U.S.A. |
| Author Unknown, “Standard RAID Levels,” Wikipedia, the Free Encyclopedia, last updated Jul. 18, 2014, 7 pages; http://en.wikipedia.org/wiki/Standard_RAID_levels. |
| Author Unknown, “Storage Infrastructure for the Cloud,” Solution Brief, © 2012, 3 pages; coraid, Redwood City, California, U.S.A. |
| Author Unknown, “Storage Area Network—NPIV: Emulex Virtual HBA and Brocade, Proven Interoperability and Proven Solution,” Technical Brief, Apr. 2008, 4 pages, Emulex and Brocade Communications Systems. |
| Author Unknown, “The Fundamentals of Software-Defined Storage, Simplicity at Scale for Cloud-Architectures” Solution Brief, 2013, 3 pages; Coraid, Redwood City, California, U.S.A. |
| Author Unknown, “VirtualWisdom® SAN Performance Probe Family Models: Probe FC8, HD, and HD48,” Virtual Instruments Data Sheet, Apr. 2014 Virtual Instruments. All Rights Reserved; 4 pages. |
| Author Unknown, “Xgig Analyzer: Quick Start Feature Guide 4.0,” Feb. 2008, 24 pages, Finisar Corporation, Sunnyvale, California, U.S.A. |
| Author Unknown, “Sun Storage Common Array Manager Installation and Setup Guide,” Software Installation and Setup Guide Version 6.7.x 821-1362-10, Appendix D: Configuring In-Band Management, Sun Oracle; retrieved and printed Sep. 12, 2013, 15 pages. |
| Author Unknown, “Vblock Solution for SAP: Simplified Provisioning for Operation Efficiency,” VCE White Paper, VCE—The Virtual Computing Environment Company, Aug. 2011, 11 pages. |
| Berman, Stuart, et al., “Start-Up Jeda Networks in Software Defined Storage Network Technology,” Press Release, Feb. 25, 2013, 2 pages, http://www.storagenewsletter.com/news/startups/jeda-networks. |
| Borovick, Lucinda, et al., “White Paper, Architecting the Network for the Cloud,” IDC Analyze the Future, Jan. 2011, pp. 1-8. |
| Chakrabarti, Kaushik, et al., “Locally Adaptive Dimensionality Reduction for Indexing Large Time Series Databases,” ACM Transactions on Database Systems, vol. 27, No. 2, Jun. 2009, pp. 188-228. |
| Chandola, Varun, et al., “A Gaussian Process Based Online Change Detection Algorithm for Monitoring Periodic Time Series,” Proceedings of the Eleventh SIAM International Conference on Data Mining, SDM 2011, Apr. 28-30, 2011, 12 pages. |
| Cisco Systems, Inc. “N-Port Virtualization in the Data Center,” Cisco White Paper, Cisco Systems, Inc., Mar. 2008, 7 pages. |
| Cisco Systems, Inc., “Best Practices in Deploying Cisco Nexus 1000V Series Switches on Cisco UCS B and C Series Cisco UCS Manager Servers,” White Paper, Cisco Systems, Inc., Apr. 2011, 36 pages. |
| Cisco Systems, Inc., “Cisco Prime Data Center Network Manager 6.1,” At-A-Glance, © 2012, 3 pages. |
| Cisco Systems, Inc., “Cisco Prime Data Center Network Manager,” Release 6.1 Data Sheet, © 2012, 10 pages. |
| Cisco Systems, Inc., “Cisco Unified Network Services: Overcome Obstacles to Cloud-Ready Deployments,” White Paper, Cisco Systems, Inc., Jan. 2011, 6 pages. |
| Clarke, Alan, et al., “Open Data Center Alliance Usage: Virtual Machine (VM) Interoperability in a Hybrid Cloud Environment Rev. 1.2,” Open Data Center Alliance, Inc., 2013, pp. 1-18. |
| Cummings, Roger, et al., Fibre Channel—Fabric Generic Requirements (FC-FG), Dec. 4, 1996, 33 pages, American National Standards Institute, Inc., New York, New York, U.S.A. |
| Farber, Franz, et al. “An In-Memory Database System for Multi-Tenant Applications,” Proceedings of 14th Business, Technology and Web (BTW) Conference on Database Systems for Business, Technology, and Web, Feb. 28 to Mar. 4, 2011, 17 pages, University of Kaiserslautern, Germany. |
| Guo, Chang Jie, et al., “IBM Resarch Report: Data Integration and Composite Business Services, Part 3, Building a Multi-Tenant Data Tier with with [sic] Access Control and Security,” RC24426 (C0711-037), Nov. 19, 2007, 20 pages, IBM. |
| Hedayat, K., et al., “A Two-Way Active Measurement Protocol (TWAMP),” Network Working Group, RFC 5357, Oct. 2008, 26 pages. |
| Horn, C., et al., “Online anomaly detection with expert system feedback in social networks,” 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 22-27, 2011, 2 pages, Prague; [Abstract only]. |
| Hosterman, Cody, et al., “Using EMC Symmetrix Storage inVMware vSph ere Environments,” Version 8.0, EMC2Techbooks, EMC Corporation; published on or about Jul. 8, 2008, 314 pages; [Retrieved and printed Sep. 12, 2013]. |
| Hu, Yuchong, et al., “Cooperative Recovery of Distributed Storage Systems from Multiple Losses with Network Coding,” University of Science & Technology of China, Feb. 2010, 9 pages. |
| Keogh, Eamonn, et al., “Dimensionality Reduction for Fast Similarity Search in Large Time Series Databases,” KAIS Long Paper submitted May 16, 2000; 19 pages. |
| Kolyshkin, Kirill, “Virtualization in Linux,” Sep. 1, 2006, pp. 1-5. |
| Kovar, Joseph F., “Startup Jeda Networks Takes SDN Approach to Storage Networks,” CRN Press Release, Feb. 22, 2013, 1 page, http://www.crn.com/240149244/printablearticle.htm. |
| Lampson, Butler, W., et al., “Crash Recovery in a Distributed Data Storage System,” Jun. 1, 1979, 28 pages. |
| Lewis, Michael E., et al., “Design of an Advanced Development Model Optical Disk-Based Redundant Array of Independent Disks (RAID) High Speed Mass Storage Subsystem,” Final Technical Report, Oct. 1997, pp. 1-211. |
| Lin, Jessica, “Finding Motifs in Time Series,” SIGKDD'02 Jul. 23-26, 2002, 11 pages, Edmonton, Alberta, Canada. |
| Linthicum, David, “VM Import could be a game changer for hybrid clouds”, InfoWorld, Dec. 23, 2010, 4 pages. |
| Long, Abraham Jr., “Modeling the Reliability of RAID Sets,” Dell Power Solutions, May 2008, 4 pages. |
| Ma, Ao, et al., “RAIDShield: Characterizing, Monitoring, and Proactively Protecting Against Disk Failures,” FAST '15, 13th USENIX Conference on File and Storage Technologies, Feb. 16-19, 2015, 17 pages, Santa Clara, California, U.S.A. |
| Mahalingam, M., et al., “Virtual eXtensible Local Area Network (VXLAN): A Framework for Overlaying Virtualized Layer 2 Networks over Layer 3 Networks,” Independent Submission, RFC 7348, Aug. 2014, 22 pages; http://www.hjp.at/doc/rfc/rfc7348.html. |
| McQuerry, Steve, “Cisco UCS M-Series Modular Servers for Cloud-Scale Workloads,” White Paper, Cisco Systems, Inc., Sep. 2014, 11 pages. |
| Monia, Charles, et al., IFCP—A Protocol for Internet Fibre Channel Networking, draft-monia-ips-ifcp-00.txt, Dec. 12, 2000, 6 pages. |
| Mueen, Abdullah, et al., “Online Discovery and Maintenance of Time Series Motifs,” KDD'10 The 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Jul. 25-28, 2010, 10 pages, Washington, DC, U.S.A. |
| Muglia, Bob, “Decoding SDN,” Jan. 14, 2013, Juniper Networks, pp. 1-7, http://forums.juniper.net/t5/The-New-Network/Decoding-SDN/ba-p/174651. |
| Murray, Joseph F., et al., “Machine Learning Methods for Predicting Failures in Hard Drives: A Multiple-Instance Application,” Journal of Machine Learning Research 6 (2005), pp. 783-816; May 2005, 34 pages. |
| Nelson, Mark, “File Verification Using CRC,” Dr. Dobb's Journal, May 1, 1992, pp. 1-18, XP055130883. |
| Pace, Alberto, “Technologies for Large Data Management in Scientific Computing,” International Journal of Modern Physics C., vol. 25, No. 2, Feb. 2014, 72 pages. |
| Pinheiro, Eduardo, et al., “Failure Trends in a Large Disk Drive Population,” FAST '07, 5th USENIX Conference on File and Storage Technologies, Feb. 13-16, 2007, 13 pages, San Jose, California, U.S.A. |
| Raginsky, Maxim, et al., “Sequential Anomaly Detection in the Presence of Noise and Limited Feedback,” arXiv:0911.2904v4 [cs.LG] Mar. 13, 2012, 19 pages. |
| Saidi, Ali G., et al., “Performance Validation of Network-Intensive Workloads on a Full-System Simulator,” Interaction between Operating System and Computer Architecture Workshop, (IOSCA 2005), Austin, Texas, Oct. 2005, 10 pages. |
| Sajassi, A., et al., “BGP MPLS Based Ethernet VPN,” Network Working Group, Oct. 18, 2014, 52 pages. |
| Sajassi, Ali, et al., “A Network Virtualization Overlay Solution using EVPN,” L2VPN Workgroup, Nov. 10, 2014, 24 pages; http://tools.ietf.org/pdf/draft-ietf-bess-evpn-overlay-00.pdf. |
| Sajassi, Ali, et al., “Integrated Routing and Bridging in EVPN,” L2VPN Workgroup, Nov. 11, 2014, 26 pages; http://tools.ietf.org/pdf/draft-ietf-bes-evpn-inter-subnet-forwarding-00.pdf. |
| Schroeder, Bianca, et al., “Disk failures in the real world: What does an MTTF of 1,000,000 hours mean to you?” FAST '07: 5th USENIX Conference on File and Storage Technologies, Feb. 13-16, 2007, 16 pages, San Jose, California, U.S.A. |
| Shue, David, et al., “Performance Isolation and Fairness for Multi-Tenant Cloud Storage,” USENIX Association, 10th USENIX Symposium on Operating Systems Design Implementation (OSDI '12), 2012, 14 pages; https://www.usenix.org/system/files/conference/osdi12/osdi12-final-215.pdf. |
| Staimer, Marc, “Inside Cisco Systems' Unified Computing System,” Dragon Slayer Consulting, Jul. 2009, 5 pages. |
| Swami, Vijay, “Simplifying SAN Management for VMWare Boot from SAN, Utilizing Cisco UCS and Palo,” posted May 31, 2011, 6 pages. |
| Tate, Jon, et al., “Introduction to Storage Area Networks and System Networking,” Dec. 2017, 302 pages, ibm.com/redbooks. |
| Vuppala, Vibhavasu, et al., “Layer-3 Switching Using Virtual Network Ports,” Computer Communications and Networks, 1999, Proceedings, Eight International Conference in Boston, MA, USA, Oct. 11-13, 1999, Piscataway, NJ, USA, IEEE, ISBN: 0-7803-5794-9, pp. 642-648. |
| Weil, Sage A., “CEPH: Reliable, Scalable, and High-Performance Distributed Storage,” Dec. 2007, 239 pages, University of California, Santa Cruz. |
| Zhou, Zihan, et al., “Stable Principal Component Pursuit,” arXiv:1001.2363v1 [cs.IT], Jan. 14, 2010, 5 pages. |
| Zhu, Yunfeng, et al., “A Cost-based Heterogeneous Recovery Scheme for Distributed Storage Systems with RAID-6 Codes,” University of Science & Technology of China, 2012, 12 pages. |
| Aweya, James, et al., “Multi-level active queue management with dynamic thresholds,” Elsevier, Computer Communications 25 (2002) pp. 756-771. |
| Petersen, Chris, “Introducing Lightning: A flexible NVMe JBOF,” Mar. 9, 2016, 6 pages. |
| Number | Date | Country | |
|---|---|---|---|
| 20170371558 A1 | Dec 2017 | US |
