Patent Application
20070260732
This invention generally relates to data storage management systems and methods thereof and, more particularly, to automatic, user load balancing systems and methods thereof.
In the emerging Software as a Service (“SaaS”) market, the majority of the processing occurs at the server and the data is centralized to take advantage of the anywhere, anytime connectivity that the Internet provides. The power of the connectivity to and sharing of information is leading to massively scalable applications that support hundreds of thousands, up to hundreds of millions of users.
Massively scalable applications are creating many new challenges in managing the users and data in an automated fashion. The ability to manage user load is particularly critical in data-heavy applications; such as email, file storage, and online backup. A user load is a load on a given server or servers using one or more services and an overall load includes the utilization of processor, memory, I/O reads, writes and transactions per second, network, disk space, power, and application or applications. To manage overall loads, currently administrators are forced to manually move these data heavy user loads between servers which is time consuming and efficient. Additionally, these manual moves of user loads can lead to interruption in service to users.
Load balancers have been developed, but they reside at the edge of the network in front of application servers and are only available to take and split incoming traffic data between servers based on low level metrics. These low level metrics consist of factors, such as the number of network connections on a server or how fast a server is responding to HTRP requests, and are unrelated to either user loads or overall loads on the servers. These prior load balancers work well for web servers where all of the data on each server is identical or “stateless,” but do not work in applications where each user has unique data, such as in the emerging SaaS market or Web 2.0, where the data on each server is unique to each user or “stateful” and thus is dynamically changing for each user. Additionally, these load balancers do not work well were each of the user loads utilize different applications.
A method for user load balancing in accordance with embodiments of the present invention includes identifying when an overall load in at least one of two or more servers exceeds one or more thresholds that are related to one or more user loads. One or more of the user loads in the identified overall load in one of the servers are transferred to one or more of the other servers when the one or more thresholds is exceeded.
A computer readable medium having stored thereon instructions for user load balancing in accordance with embodiments of the present invention identifying when an overall load in at least one of two or more servers exceeds one or more thresholds that are related to one or more user loads. One or more of the user loads in the identified overall load in one of the servers are transferred to one or more of the other servers when the one or more thresholds is exceeded.
A user load balancing system in accordance with embodiments of the present invention includes a threshold system and a transfer system. The threshold system identifies when an overall load in at least one of two or more servers exceeds one or more thresholds. Each of the overall loads comprises one or more user loads and each of the one or more thresholds is based on the user loads. The transfer system initiates a transfer of one or more of the user loads in the identified overall load in one of the servers to one or more of the other servers when the one or more thresholds is exceeded.
The present invention provides a number of advantages including providing an effective and automatic method and system for balancing user loads, such as stateful data comprising user emails, between two or more servers, such as storage servers, partitions of a centralized storage, or application servers. Additionally, with the present invention the balancing of the user loads between storage servers and/or application servers is seamless to the end user and occurs without an interruption in service. Further, the present invention is able to ensure the balancing of the user loads without loss of any of the user loads. Even further, the present invention is able to distinguish between temporary heavy usage which would not require intervention, and might even hinder performance, and more persistent heavy usage requiring user load balancing.
A system 10 with a user load balancing system 12 in accordance other embodiments of the present invention is illustrated in
Referring more specifically to
The user load balancing system 12 comprises a central processing unit (CPU) or processor 20, a memory 22, and an interface system 24 which are coupled together by a bus or other link 26, although the can comprise other numbers and types of each of the components and other configurations and locations for each of the components can be used. The processor 20 executes a program of stored instructions for one or more aspects of the present invention as described and illustrated herein, including the method for user load balancing, although the processor 20 could execute other types of programmed instructions. The memory 22 stores these programmed instructions for one or more aspects of the present invention as described herein, although some or all of the programmed instructions could be stored and/or executed elsewhere. A variety of different types of memory storage devices, such as a random access memory (RAM) or a read only memory (ROM) in the system or a floppy disk, hard disk, CD ROM, or other computer readable medium which is read from and/or written to by a magnetic, optical, or other reading and/or writing system that is coupled to one of the processor, can be used for the memory 22.
The interface system 24 is used to operatively couple and communicate between the user load balancing system 12 and the global directory server 14, the storage servers 16(l)-16(n), and the application servers 17(l)-17(n) via communications system 18 although other types and numbers of connections and other configurations could be used. In this particular embodiment, the communication system 18 is via TCP/IP over an Ethernet and uses industry-standard protocols including SOAP, XML, LDAP, and SNMP, although other types and numbers of communication systems, such as a direct connection, another local area network, a wide area network, Internet, modems and phone lines, e-mails, and/or wireless communication technology each having their own communications protocols, could be used.
The global directory server 14 includes a centralized master directory of locations where servers, services, and user data reside, although the global directory server can store other types of information. Any user via a service or application can query the global directory server 14 to find out where given data in a user load is located.
The global directory server 14 also includes a central processing unit (CPU) or processor, a memory, and an interface system which are coupled together by a bus or other link, although other numbers and types of each of the components and other configurations and locations for the components can be used. For example, the global directory server 14 can comprise a single server as shown or can be replicated in a series of masters and slaves. In a master/slave environment, the master global directory server will be in charge of all updates and the slave global directory servers will be read servers that respond to application requests. This enables the directory in the global directory server 14 to scale.
In this particular embodiment, the processor in the global directory server 14 shown in
Each of the storage servers 16(l)-16(n) is used to store user loads comprising stateful and stateless data, although one or more of the storage servers 16(l)-16(n) could have other functions. This storage could be local or could be over SAN, NAS, iSCSI, distributed file systems, or other types of local or remote network systems. Each of the user loads stored in the storage servers 16(l)-16(n) comprises stateful data and stateless data, although other types and combinations of data can be stored, such as just user loads with just stateful data. In this particular embodiment, each of the storage servers 16(l)-16(n) includes a central processing unit (CPU) or processor, a memory, and an interface system which are coupled together by a bus or other link, although other numbers and types of each of the components and other configurations and locations for the components can be used. Additionally, one or more other types of servers can be used with or in place of storage servers 16(l)-16(n), such as network area storage or one or more application servers.
The processor in each of the storage servers 16(l)-16(n) executes a program of stored instructions for one or more aspects of the present invention as described herein, including instructions for storing stateful and stateless data and for transferring data between storage servers 16(l)-16(n) and/or application servers 17(l)-17(n). The memory stores these programmed instructions for one or more aspects of the present invention as described herein, although some or all of the programmed instructions could be stored and/or executed elsewhere. A variety of different types of memory storage devices, such as a random access memory (RAM) or a read only memory (ROM) in the system or a floppy disk, hard disk, CD ROM, or other computer readable medium which is read from and/or written to by a magnetic, optical, or other reading and/or writing system that is coupled to the processor, can be used for the memory in the management server system. The interface system in each of the storage servers 16(l)-16(n) is used to operatively couple and communicate between the storage servers 16(l)-16(n) and the user load balancing system 12, the global directory server 14, and the application servers 17(l)-17(n), although other types and connections could be used.
Each of the application servers 17(l)-17(n) has one or more user loads relating to the execution of one or more applications, such as email or one or more other applications, although one or more of the application servers 17(l)-17(n) could have other functions and other types and numbers of systems could be used. Each of the application servers 17(l)-17(n) also includes a central processing unit (CPU) or processor, a memory, and an interface system which are coupled together by a bus or other link, although other numbers and types of each of the components and other configurations and locations for the components can be used.
The processor in each of the application servers 17(l)-17(n) executes a program of stored instructions for one or more aspects of the present invention as described herein, including instructions for applications, such as email and/or one or more other applications, and for transferring data between application servers 17(l)-17(n) and/or storage servers 16(l)-16(n). The memory stores these programmed instructions for one or more aspects of the present invention as described herein, although some or all of the programmed instructions could be stored and/or executed elsewhere, such as in one or memories of provider systems. A variety of different types of memory storage devices, such as a random access memory (RAM) or a read only memory (ROM) in the system or a floppy disk, hard disk, CD ROM, or other computer readable medium which is read from and/or written to by a magnetic, optical, or other reading and/or writing system that is coupled to the processor, can be used for the memory in the management server system. The interface system in each of the storage servers 17(l)-17(n) is used to operatively couple and communicate between the storage servers 17(l)-17(n) and the predictive capacity planning system 12, the global directory server system 14, and storage servers 16(l)-16(n), although other types and connections could be used.
Although an example of embodiments of the user load balancing system 12, the global directory server 14, the storage servers 16(l)-16(n), application servers 17(l)-17(n) is described and illustrated herein, each of the user load balancing system 12, the global directory server 14, the storage servers 16(l)-16(n), and application servers 17(l)-17(n) of the present invention could be implemented on any suitable computer system or computing device. It is to be understood that the devices and systems of the exemplary embodiments are for exemplary purposes, as many variations of the specific hardware and software used to implement the exemplary embodiments are possible, as will be appreciated by those skilled in the relevant art(s).
Furthermore, each of the systems of the present invention may be conveniently implemented using one or more general purpose computer systems, microprocessors, digital signal processors, micro-controllers, and the like, programmed according to the teachings of the present invention as described and illustrated herein, as will be appreciated by those skilled in the computer and software arts.
In addition, two or more computing systems or devices can be substituted for any one of the systems in any embodiment of the present invention. Accordingly, principles and advantages of distributed processing, such as redundancy, replication, and the like, also can be implemented, as desired, to increase the robustness and performance the devices and systems of the exemplary embodiments. The present invention may also be implemented on computer systems that extend across any network using any suitable interface mechanisms and communications technologies including, for example telecommunications in any suitable form (e.g., voice, modem, and the like), wireless communications media, wireless communications networks, cellular communications networks, G3 communications networks, Public Switched Telephone Network (PSTNs), Packet Data Networks (PDNs), the Internet, intranets, a combination thereof, and the like.
The present invention may also be embodied as a computer readable medium having instructions stored thereon for user load balancing as described herein, which when executed by a processor, cause the processor to carry out the steps necessary to implement the methods of the present invention as described and illustrated herein.
The operation of the user load balancing system 12 in a system 10 will now be described with reference to
By way of example only, the types thresholds being monitored by the user load balancing system include: disk space utilization (measured in both total Gigabytes (GB) of usable storage as well as measured in a percentage of 100%); RAM utilization (measured in percentages of 100%); I/O load (measured by the number of I/O reads per second, the number of I/O writes per second and the number of transactions per second; and network load (measured as a percentage of 100%), although other types and numbers of thresholds related to user loads and/or overall loads can be monitored. Additionally, the user load balancing system 12 can be configured by to allow an administrator to enter and/or adjust one or more of the particular thresholds as needed or desired for the particular application.
More specifically, in these embodiments the user load balancing system 12 monitors for short and long term threshold trends in the overall loads on the storage servers 16(l)-16(n) and application servers 17(l)-17(n). These short and long term threshold trends or thresholds are entered by an administrator in the user load balancing system, although other configurations can be used, such as having the short and long term thresholds stored and retrieved from memory in the user load balancing system 12.
By way of example only, short term thresholds monitored by the user load balancing system 12 include: processor utilization at greater than 90% for longer than five minutes; RAM utilization at greater than 90% for longer than 20 minutes; network load higher than 90% for longer than five minutes; and I/O read/writes at greater than 90% of capacity for longer than 30 minutes, although these are exemplary and again other types and numbers of short term threshold trends can be monitored. Additionally, by way of example only, long term thresholds monitored by the user load balancing system 12 include: processor load greater than 90% average for more than 180 minutes; disk space utilization greater than 95%; RAM utilization at greater than 90% average for more than 300 minutes; and I/O reads/writes in excess of 85% of capacity for more than 120 minutes, although these are exemplary and again other types and numbers of long term threshold trends can be monitored.
In step 32, the user load balancing system 12 determines if at least one of the thresholds has been exceeded, although other numbers and types of thresholds could be established and checked. If none of the thresholds has been exceeded, then the No branch is taken back to step 30 where the overall load on each of the storage servers 16(l)-16(n) and on each of the application servers 17(l)-17(n) is monitored by the user load balancing system 12. If one or more of the thresholds has been exceeded, then the on or more overall loads which exceeded the threshold are identified and the Yes branch is taken to step 34.
In step 34, the user load balancing system 12 determines which of the one or more user loads in the identified overall load or loads in one or more of the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) to move to one or more of the other storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) to even out the overall load among all of the available storage servers 16(l)-16(n) and/or application servers 17(l)-17(n), although the user load balancing system 12 can use other manners for identifying which of the user loads to move.
In step 36, the user load balancing system 12 determines an order for moving the user loads identified for transfer in step 34 based on one or more characteristics. By way of example only, these characteristics include identifying which of the user loads: is currently inactive; caused the identified overall load in at least one of the storage server or application server to exceed the threshold; is the largest in the identified overall load; and has the last login date to an application, although other numbers and types of characteristics could be used for determining the order of transfer, such as identifying the services users are using. The particular parameters for determining the order based on these characteristics can be set by the administrator. By way of example only: the user load not currently logged on could be moved first; a single user load which caused the threshold to be exceeded could be moved first; the largest user load could be moved first; or the user load with the oldest login date could be moved first.
Next, in step 38 the user load balancing system 12 determines which of the one or more other storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) to move each of the determined one or more user loads, although the user loads could be moved to other types of servers, such as application servers. By way of example only, the user load balancing system 12 may determine which of the other storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) has the lowest average utilization of at least one of processor space and memory space, although other utilizations could be examined, such as I/O reads, writes, and transactions per second, network, disk space, power, and applications. Additionally by way of example only the user load balancing system 12 may determine which of the one or more other storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) need to receive one or more of the user loads from the identified overall load in the at least one of the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) to balance the stored load between the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n). As part of this process, the user load balancing system 12 monitors for the addition of any storage server and/or application server that could be considered for transferring some of the user load or user loads. The user load balancing system 12 could actively scan for the addition of new storage servers and/or application servers, although other manners for identifying new storage servers and/or application servers could be used, such as by a notification input by an administrator.
Next, in step 40 the user load balancing system 12 initiates the transfer of the one or more user loads which have been identified and the transfer takes place directly between the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n), although other manners for initiating the transfer and for transferring the user loads could be used.
In step 42, the user load balancing system 12 optionally determines if the transferred one or more user loads match identically with the original one or more user loads, although other types of verifications may be conducted. By way of example only, the user load balancing system 12 performs a checksum, although other methods for assuring an accurate transfer of data can be used or no check may be performed. If the one or more of the user loads have not been correctly copied, then the No branch is taken back to step 40 where the user load balancing system 12 transfers the one or more of the user loads again. If the one or more identified user loads have been correctly transferred, then the Yes branch is taken to step 44.
In step 44, since stateful data, such as email, may continue to flow to the user load being transferred, following the transfer the user load balancing system 12 checks for any incremental updates in the original user loads which may have occurred following the transferring, although other methods for handling incremental updates can be used. For example, when a user load is designated for transfer, the user load could be placed in a queue and only a read only version would be accessible by a user until the transfer of the user load has been completed. Alternatively, during the transfer the system 10 could prevent access to the user load being transferred until the transfer is completed. Returning to step 44, if there are no incremental updates, then the No branch is taken to step 50. If there are incremental updates, then the Yes branch is taken to step 46. In step 46, the user load balancing system 12 transfers the incremental updates into the transferred user loads.
In step 48, the user load balancing system 12 determines if the incremental updates have been correctly transferred, although other types of verifications may be conducted. If the transfer is incorrect, then the No branch is taken back to step 46 where the user load balancing system 12 attempts the transfer of the incremental updates again. Additionally, the user load balancing system 12 may provide a notification that the transfer has failed, although the user load balancing system 12 can provide notification relating to the success or failure of other events or not provide any notification. Further, the user load balancing system 12 can include a history of transactions which have taken place that can be accessed by an administrator, if desired. If the transfer is correct, then the Yes branch is taken to step 50.
In step 50, the user load balancing system 12 updates the directory with the new address or addresses for the user loads. In step 50, the user load balancing system 12 deletes the original user loads in the one or more of the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) that exceeded the threshold. This frees up resources and should place the overall load in the identified one or more of the storage servers 16(l)-16(n) and/or application servers 17(l)-17(n) below the one or more thresholds. Following step 52, the user load balancing system 12 returns to step 30 for ongoing monitoring of the overall loads in storage servers 16(l)-16(n) and application servers 17(l)-17(n).
Accordingly, the present invention provides an efficient system for balancing user loads comprising stateful data between servers. Additionally, the present invention is able to balance the user loads without requiring any downtime to move user loads between the servers. Further, the present invention is able to ensure the balancing of the user loads without loss of any of the user loads. Even further, as illustrated above the present invention is able to distinguish between temporary heavy usage which would not require intervention and more persistent heavy usage requiring user load balancing through the use of short and long term.
Having thus described the basic concept of the invention, it will be rather apparent to those skilled in the art that the foregoing detailed disclosure is intended to be presented by way of example only, and is not limiting. Various alterations, improvements, and modifications will occur and are intended to those skilled in the art, though not expressly stated herein. These alterations, improvements, and modifications are intended to be suggested hereby, and are within the spirit and scope of the invention. Additionally, the recited order of processing elements or sequences, or the use of numbers, letters, or other designations therefore, is not intended to limit the claimed processes to any order except as may be specified in the claims. Accordingly, the invention is limited only by the following claims and equivalents thereto.
