Embodiments of the present invention are directed generally toward computer networks and, more particularly, to monitoring and diagnostics of remote electronic networks.
Home networks are growing exponentially in their complexity. We will shortly see home networks where the computer is merely one participant among many types of devices. The number of PCs attached to home networks will be dwarfed by other devices such as televisions and DVRs that are connected directly to the home network.
Except in rare cases, the user in the home will never be an IT expert or even have more than a very limited understanding of what a network does. As more and more devices use the home network, the burden of managing the network will rapidly grow beyond the capabilities of most home users. From the user's perspective, very simple tasks will seem as though they do not work, when in reality, they do work but are being hampered by a non-optimal network configuration or by events in the network they do not understand yet could control if they were armed with the appropriate information.
Consider the network environment illustrated in
This family has a daughter who has just received a laptop computer 170 for her birthday. As illustrated in
The daughter installs a particularly “chatty” peer-to-peer (p2p) file sharing application on her computer 170, and suddenly this high bandwidth application makes resources available from the connection to the Internet 107 relatively scarce.
Mom logs on to her computer 155 and starts a video session with Grandma. It is likely that Mom's chat session will begin to experience lost packets and “jitter” as she unknowingly competes with her daughter for bandwidth. Mom's experience will be degraded. She might have some vague notion that things seemed to get worse at some point after they installed the router 165. She might blame the router 165, she might blame the broadband service provider (BSP), she may simply live with the degraded experience and think there's nothing she can do about it, or she may simply give up and decide that video chat is not a viable option.
It's unlikely that, with no experience in networking, Mom will even begin to understand that her experience is being degraded as her network packets compete with her daughter's packets for a relatively scarce resource.
Furthermore, from the perspective of the BSP, they would rather that the scarce resource of the cable feed to this family's neighborhood not be saturated with p2p file sharing traffic.
Home network monitoring is very limited at present. For the most part it is limited to watching whether or not the Internet connection is active or whether link layer connectivity is functional on the local computer's network adapter. As the number of devices on the network increases, these limitations must be addressed.
An embodiment of the invention includes a system for optimizing a network having a plurality of network devices. The system includes a network monitoring tool simultaneously executable on at least one of the network devices and configured to diagnose at least one condition of the network. The network monitoring tool includes an information collection module configured to collect an information set relating to performance of each of the network devices of the network, an action module configured to execute a predetermined action in response to at least a portion of the information set, the action relating to a diagnosed condition of the network, and an information transmission module configured to transmit via a wide-area network at least a portion of the information set to an electronic device remote from the network. The system further includes a configuring tool executable on a server computing device. The configuring tool includes an information receiving module configured to receive the information-set portion, and a programming module configured to configure via the wide-area network the network monitoring tool based on the received collected information.
Preferred and alternative embodiments of the present invention are described in detail below with reference to the following figures:
An embodiment of the invention includes a home-network (client-side) monitoring system that provides for centralized configuration, troubleshooting and diagnosis of the home network environment. The system may be controlled via a third party such as a BSP responsible for the management of multiple client home networks. The system can upload data into a provider-managed (server-side) repository allowing the provider to analyze and reconfigure the customer's monitoring environment. Furthermore, diagnostics and performance data can be automatically fed by the client-side monitoring system into the provider's repository allowing for simplified troubleshooting steps in the event the customer calls the provider's technical support.
Data uploaded to the centralized system (e.g., the BSP) may be simple raw metric information or events that are generated in response to analysis of the raw metric information. For example, an embodiment might collect raw network usage statistics from each device in a home and upload that information every hour. Furthermore, there might be rules that generate an event whenever the network utilization from any one device crosses a threshold. Those events may be uploaded in addition to the raw data that is uploaded periodically. The raw metric information potentially provides a customer service representative (CSR) with valuable baseline (i.e., normal state) information for the customer, and events generated from the analysis of that information can provide information about specific points in time where the behavior of the network deviated from what is normal.
In an embodiment, the home-network monitoring system can be dynamically configured from a central (e.g., server-side) location as a basis for discovery. An embodiment of the invention described within includes a system that allows for dynamic and central configuration of a home-network monitoring system managed either locally or from a central location with no home user intervention. The client-side monitoring system can automatically push data into the repository and accept new rule and diagnostic configurations from the repository that modify the functionality of the client-side monitoring system. The data may be collected from individual computing devices in the home network. An embodiment can process data about the home network from a variety of device types using potentially many different protocols.
As earlier alluded to, diagnostic and performance information uploaded into the repository may be available to CSRs associated with the BSP for the following purposes:
a. Assisting with customer service calls. The CSRs can use the information uploaded from a customer site to help diagnose problems in that customer's network environment.
b. Analyzing data from all customers to provide new rules and diagnostics to be applied across the entire customer base.
Problem signatures that arise from the analysis of data uploaded to the repository can be used for developing new rules that identify such signatures. This allows the BSP to fix the problem remotely or provide information to the home network customer prior so as to potentially prevent the need to call a CSR for assistance.
An embodiment allows an entity, external to a home network, to control the rules governing monitoring data collected in the home network, such network having no standardized mechanism for collecting such data.
As will be more fully discussed hereinafter, the system allows the BSP to provision the collection of new data through simple modifications of a configuration file or through the addition of new configuration files associated with the client-side monitoring system. These modifications/additions can be facilitated through a push model whereby the BSP pushes the data to the home network or a pull model whereby the home network periodically checks for downloadable changes to be applied. The home network may check for such changes in response to a triggered rule.
An embodiment allows the BSP to provision new rules that perform analysis on arbitrary metrics through the modification of existing configuration files or the addition of new configuration files. Additionally, the BSP can provision new diagnostic operations to the home network through the modification of existing configuration files or the addition of new configuration files.
An embodiment is flexible enough to enable the BSP to push rules to a customer only when those rules are relevant to that particular customer's situation. Thresholds may be automatically tuned by software or tuned by a CSR to accommodate the user's environment.
Embodiments of the invention may incorporate or otherwise utilize a dedicated software application tool for managing small networks and described in detail in U.S. Provisional Patent Application No. 60/634,432, filed Dec. 7, 2004, entitled “Network Management” and naming Steve Bush et al. as inventors, and U.S. patent application Ser. No. 11/297,809, filed on Dec. 7, 2005, entitled “Network Management” and naming Steve Bush et al. as inventors, which applications, along with U.S. Provisional Patent Application No. 60/789,522, filed Apr. 4, 2006, entitled “Network Management,” U.S. patent application Ser. No. 10/916,642, filed on Aug. 10, 2004, entitled “Service Licensing And Maintenance For Networks,” U.S. patent application Ser. No. 11/457,783, filed on Jul. 14, 2006, entitled “Network Device Management,” and U.S. patent application Ser. No. 11/457,763, filed on Jul. 14, 2006, entitled “Network Device Setup Utility,” are incorporated entirely herein by reference.
As previously noted, various embodiments of the invention may be employed in connection with a small network.
Each of these networked devices 103 communicates, either directly or indirectly, with a gateway or router device 105. In turn, the gateway device 105 typically will communicate with an external device or network. An external network may be another private network, or it may be a public network, such as the Internet 107. Thus, a gateway device is a device that can steer electronic data from one network to another network. Typically, a gateway device serves as a node on two incompatible networks (i.e., networks that use different communication protocol formats) and it can convert data from one network's communication protocol format into the other network's communication protocol format. As used herein, the term “small network” refers to a network made up of networked devices that each employ the same network address to communicate with the same gateway device, together with the gateway device itself.
The network devices 103 may be connected to the gateway device 105 using any suitable communication medium. For example, in the illustrated network 101, the desktop computers 103B are connected to the gateway device 105 through a hard-wired connection 109A (such as an Ethernet cable), while the laptop computer 103A is connected to the gateway device 105 through a IEEE 802.11 wireless connection 109B and the personal digital assistant 103C is connected to the gateway device 105 through a Bluetooth wireless connection 109C.
It should be appreciated that, as used throughout this application, the term “connect” and its derivatives (e.g., connection, connected, connects) includes both direct and indirect connections. Thus, with the network illustrated in
Typically, the gateway device 105 will be a router. As will be appreciated by those of ordinary skill in the art, a router routes data packets from the networked devices 103 to an external device or network. With some networks, however, the gateway device 105 alternately may be a computer performing router functions, a hub, a bridge, or “layer-3” switch. As will also be appreciated by those of ordinary skill in the art, the computing devices or “nodes” making up the network 101 can communicate with the gateway device 105 using one or more defined communication protocols, such as the Transmission Control Protocol (TCP) and the Internet Protocol (IP).
With these communication protocols, each computing device 103 and gateway device 105 in the network 101 can be assigned a logical address. For example, if the network 101 is connected to the Internet 107 through an Internet service provider, the Internet service provider can assign the gateway device 105 a logical Internet Protocol (IP) address. The Internet service provider may also provide the gateway device 105 with a block of logical Internet Protocol (IP) addresses for the gateway device 105 to reassign to each network device 103. Alternatively, the gateway device 105 can itself assign a range of logical Internet Protocol (IP) addresses to each network device 103, and then use a translation operation (e.g., a Network Address Translation (NAT) operation) to route data packets that it receives to the appropriate network device 103. This type of logical address typically is unrelated to the particular computing device to which it is assigned. Instead, a logical address identifies the relationship of that computing device to other computing devices in the network.
In addition to a logical address, each network device typically can also have a physical address. For example, most computing devices capable of communicating over a network, including routers, employ a network adapter with a media access control (MAC) address. This type of physical address is assigned to a network adapter according to standards (referred to as Project 802 or just 802 standards, which are incorporated entirely herein by reference) set forth by the Institute of Electrical and Electronic Engineers (IEEE). More particularly, these standards define a 48-bit and 64-bit physical address format for network devices. The first 14 bits of the address are assigned by the IEEE Registration Authority, and uniquely identify the manufacturer of the network adapter. The remaining bits are then assigned by the manufacturer to uniquely identify each network adapter produced by the manufacturer. Consequently, the physical address of a network adapter is unique across all networks unless manually changed by the user. The physical address is unique to the network adapter, and is independent of a computing device's relationship to other computing devices in a network. Thus, the physical address does not change over time or between uses in different networks.
A network may include both virtual devices and physical devices. Physical network devices can then include both computer devices and computing appliance devices. A “computer” may generally be characterized as a device that can be programmed to perform a number of different, unrelated functions. Examples of computers can thus include programmable personal computers, such as desktop computers and laptop computers. In addition, programmable media-purposed computers (e.g., “media adapters and servers”), network attached storage devices, programmable entertainment-purposed computers (e.g., video game consoles), some programmable personal digital assistants and some telephones (such as wireless “smart” telephones) may be characterized as computers in a network. A “computing appliance” then may generally be characterized as a device that is limited to primarily performing only specific functions. Examples of a computing appliance may thus include, for example, printers, cameras, telephones that exchange voice information in data packets (sometimes generically referred to as “Voice over Internet Protocol (VoIP) telephones or telephone adapters), digital video recorders, televisions, voice over Internet protocol (VoIP) adapters, print servers, media adapters, media servers, photo frames, data storage servers, routers, bridges and wireless access points.
As will be appreciated by those of ordinary skill in the art, there may be no clear defining line between “computer” network devices and “computing appliance” network devices in a network. For example, a sophisticated print server may be programmable to additionally or alternately function as a data storage server, while a programmable media-purposed computer or programmable personal digital assistant may have restricted functionality due to limited memory, input devices or output devices. Accordingly, as used herein, the term “computer” can refer to any network device that is capable of implementing a network management tool according to one or more aspects of the invention, such as a personal programmable computer. The term “computer appliance” then can refer to a network device that typically cannot implement a network management tool according to at least one aspect of the invention without additional augmentation. The term “computing device” is then used herein to include both computers and computing appliances.
With conventional networks located in a home, small office or other local environment, a network management tool according to various aspects of the invention can be implemented on a programmable personal computer, such as a desktop or laptop computer. A general description of this type of computer will therefore now be described.
An illustrative example of such a computer 201 as may be present in the network 101 described above is illustrated in
The processing unit 205 and the system memory 207 are connected, either directly or indirectly, through a bus 213 or alternate communication structure to one or more peripheral devices. For example, the processing unit 205 or the system memory 207 may be directly or indirectly connected to additional memory storage, such as the hard disk drive 215, the removable magnetic disk drive 217, the optical disk drive 219, and the flash memory card 221. The processing unit 205 and the system memory 207 also may be directly or indirectly connected to one or more input devices 223 and one or more output devices 225. The input devices 223 may include, for example, a keyboard, touch screen, a remote control pad, a pointing device (such as a mouse, touchpad, stylus, trackball, or joystick), a scanner, a camera or a microphone. The output devices 225 may include, for example, a monitor display, television, printer, stereo, or speakers.
Still further, the computing unit 203 can be directly or indirectly connected to one or more network interfaces 227 for communicating with a network. This type of network interface 227, also sometimes referred to as a network adapter or network interface card (NIC), translates data and control signals from the computing unit 203 into network messages according to a communication protocol, such as the Transmission Control Protocol (TCP), the Internet Protocol (IP), and the User Datagram Protocol (UDP). These protocols are well known in the art, and thus will not be described here in more detail. An interface 227 may employ any suitable connection agent for connecting to a network, including, for example, a wireless transceiver, a power line adapter, a modem, or an Ethernet connection.
It should be appreciated that one or more of these peripheral devices may be housed with the computing unit 203 and bus 213. Alternately or additionally, one or more of these peripheral devices may be housed separately from the computing unit 203 and bus 213, and then connected (either directly or indirectly) to the bus 213. Also, it should be appreciated that both computers and computing appliances may include any of the components illustrated in
It should be noted that, while a general description of a programmable personal computer was provided above, various aspects of the invention may be implemented on any desired device capable of supporting embodiments of the invention. For example, with some aspects of the invention, the network management tool may be implemented on special purposed programmable computers, such as a programmable media or entertainment-purposed computers, or personal digital assistants. Accordingly, the above description of a programmable personal computer should be understood as illustrative rather than limiting.
A computing appliance may have any combination of the components of the computer 201 discussed above. More typically, however, a computing appliance can be simpler to optimize the performance of a specific function, and thus may have only a subset of these components. For example, a computing appliance may have only a computing unit 203, an input device 223 or an output device 225, and a network interface 227. As will be apparent from the following description, however, a computing appliance will have sufficient computing resources to implement a desired embodiment of the invention in order to provide information to or receive information from a client operating on a separate computing device.
As shown in
As will be discussed more fully hereinafter, or as otherwise discussed in the patent applications incorporated by reference herein, one or more instantiations and/or components of a network monitoring tool 301 implemented on at least one computing device (not shown in
The tool 301 includes a set of configuration files 601. In an embodiment, the configuration file set includes data-set definitions 603, metric definitions 605, rule definitions 607, action definitions 609, target definitions 611, and state definitions 608. The updating and/or organization of the configuration files 601 may be controlled by a configuration manager component 633.
As will be discussed more fully hereinafter, the configuration file set 601 may be employed by the tool 301 in performing the following functions:
The tool 301 may be driven by a generic, easily extended and modified data file(s). For example, extensible markup language (XML) may be used to define this type of input. In an embodiment, the tool 301 may accommodate dynamically added and configured modules that can provide data from arbitrary sources. That is, the collection of data from specific hardware and/or computer systems of the network 101 may be implemented by small modules that provide data to the server 407 as tables of metrics as described above. The XML schema supports the creation and implementation of dynamically defined providers of data that can be added to the tool 301 on the fly.
Additionally, an embodiment may employ modules that include configuration data specific to the module and embedded within the module itself. For example, a new module might require configuration information that is specifically unique to it. The tool 301 supports the ability to provide that data such that it coexists with the rest of the monitoring configuration information defined in the tool.
Each state may have a set of one or more triggers that change the active state to one of the other states defined for the state machine 626. A trigger may be a rule as defined previously (e.g., a threshold is violated) or a period of time (e.g., execute this trigger after 20 seconds). When a trigger fires, the state machine changes the active state to a new state as defined for the firing trigger. When a state is entered it may invoke an action either immediately or after some period of time has passed.
An example of how a state machine 626 might be used is as follows: If only one person is using the network, it doesn't matter how much bandwidth they use. If two people are using the network, however, they are sharing a scarce resource. Restated, a computer using lots of bandwidth is only relevant if another computer is also attempting to use lots of bandwidth. As such, a state machine running on computer 1 may go into a “busy” state while that computer is busy or an “other busy” state when computer 2 is busy. Neither of those states represents a situation the user cares about. Consider computer 2 becoming busy (a trigger generated by the singular event of computer 2 becoming busy) while computer 1 is in the “busy” state or computer 1 becoming busy (a trigger generated by the singular event of computer 1 becoming busy) while in the “other busy” state. Either of these triggers may result in a state transition from the associated state to a state that invokes an action that alerts a user that there are two computers contending for limited bandwidth.
Examples might include the following:
i) Average network utilization statistics for every computer are uploaded on a timed basis;
ii) Sustained network utilization violating a threshold on multiple computers at the same time might generate an event that is uploaded to the CSR repository;
iii) In response to Internet connectivity being lost for more than five minutes, a diagnostic might reset an HNAP enabled router.
With reference back to
The BSP providing service to the family described in the scenario is interested in tracking particular types of traffic and watching for certain types of traffic breaking a threshold. This can aid them in troubleshooting. Two events in the family's network are illustrative of how an embodiment provides value. The first is an event on the Internet that creates a scenario whereby the BSP wants to provision rules to its customers. The second is a scenario where two members of the family are simultaneously consuming large amounts of broadband bandwidth, thus impacting each other's performance.
The following sample rule file shows an exemplary implementation of a configuration file that might drive the rules engine 619 described herein to provide the BSP the tracking functionality they are looking for.
In this case, a new p2p file system “Foo” has been created and let loose on the Internet. The BSP is already tracking Kazaa traffic through a configuration file provided to customers. The new configuration file, shown below, is provided to new and existing customers by the BSP. It has been modified to track the Foo system as well as Kazaa. Existing customers will get this new configuration file when the rules engine 619 checks in to see if there are new configuration files 601 it should download.
A section by section description of the exemplary implementation follows:
The Upload Targets section describes targets to which data may be uploaded. For example, if the BSP was provisioning a customer's computer with a monitoring configuration file, this is where they may define the Web service (e.g., server 407) where data should be uploaded to. In response to the triggering of some rule, the BSP might elect to have some set of data collected and shipped up to them once or each time it is collected on some fixed interval. This could provide extra detail for their CSR personnel in the event a customer calls with a problem. These targets may later be referenced in data collection sections or rule sections of the file.
The target manager 629 is responsible for processing these targets and moving data between other computers on the 101 and a remote network service provider (e.g., server 407). The monitoring system 621 or the action handler 627 might feed data to the target manager 629 for distribution. The monitoring system 621 may feed raw metrics as they are collected and the action handler may feed results of rule evaluations that triggered one or more actions.
In this example a single upload target is defined. It is a web service for the BSP that will receive event notifications from customers when a rule triggers an action that uploads event notifications to this upload target.
The provider modules 613 are configured to collect data. The two examples given provide two very different types of collection modules. This section defines modules 613 that can be used by the basic metric collector component 617 to collect raw state values from one or more network devices 103.
The first sample, PacketSignatureProvider is configured to read input files that define a signature for a type of packet. It analyzes every packet traversing one or more network adapters 105 in the network 101 and maintains counters about packets that match a particular signature. In an embodiment, this provider is created by the BSP, though it could just as easily have been created by a provider of the rules engine 619. It is, however, illustrative, in this case, of the BSP being able to modify the system on the fly.
The second sample, PerfCounterProvider is configured to collect information from Windows Performance counters. This is a general purpose protocol that allows collection of performance information pertaining to many entities in a computer running the Windows OS. The data is broken down into groups (for example network statistics or disk statistics). Each group contains multiple items (for example total bytes into a network adapter). Each of these statistics then is collected for one or more network adapters 105.
Both of these providers 613 are configured to collect metric data from specific sources and feed it back to the monitoring system 621 in a well-defined manner. These samples define the providers 613 in a construct that is compatible with, but not limited to, Microsoft.NET managed code. An alternative embodiment could use COM.
Data set definitions 603 describe the raw metrics that are to be collected for analysis. A data set definition 603 defines parameters that drive an instance of a provider 613. Data sets 603 define what data is to be collected by the provider 613, how often it should be collected and other meta-data about the set including an ID that is later used to reference the data set. The interpretation of much of the meta-data is specific to the provider 613.
The basic metric collector 617 is configured to manage the non-provider specific information from the configuration file 601. The basic metric collector 617 is further configured to pass provider-specific information down to the provider module 613 for further configuration.
Consider the Sample:
The first data set defines the packet signature information to be collected by the tool 301. The name attribute defines a name that this data set will be known by. The second attribute provider references back to the definition of the provider 613 that is configured to provide information for the data definition that will follow. In this instance, it's the packet signature analyzer. Interval specifies how often counter metrics should be collected by the rules engine 619 and id provides a unique ID by which the data set will be referenced. The ID is concatenated to the name of the input file to form a complete id for the rules engine 619; therefore the ID need only be unique within this file.
The second data set is similar, except, in this case, the class attribute is advantageous. As specified previously, performance counters are broken down into groups. The class attribute in this case defines what group of performance counters the individual data items will be drawn from.
The provider 613 is configured to expose a table of data, in the case of the network adapter statistics, that is an extensive table with many elements. DataItems specify what values from the full set of data the provider 613 is configured to collect and that should be passed into the monitoring system 621 for analysis. For some providers 613, collecting a subset of an entire table provides performance gains over collecting the entire table.
Each data item has a “friendly name” and an instanceDescriptor that describes it to the provider 613. The instanceDescriptor is the key that the provider 613 uses to identify a row in a table of data and is specific to the type of provider 613. Of the many possible network statistics that can be collected for an adapter 105, the tool 301 may, for example, only be interested in the number of bytes and packets going in and coming out.
“Instances” may be used to describe explicit sources of data. In the case of network statistics, the tool 301 may need the name of an adapter 105. The PerfCounterProvider requires nothing more than this name to collect all statistics for an adapter 105. The “name” attribute specifies “Ethernet0” which is the Ethernet adapter in the device(s) executing the tool 301. As such, there could be multiple Ethernet adapters specified here; there is no reason it need be limited to 1.
“PacketSigStats” provides an interesting example. In this case, it may be desired that each instance represents the number of packets for a particular signature going through an adapter 105. It can then be seen that the first Instance element has the name attribute “Ethernet0” describing the same adapter seen already. The next two elements are specific to the packet signature provider. When the rules engine 619 reads the configuration file, it will pass down the XML tree below each Instance to the provider so it can further differentiate instances. Furthermore, the provider is allowed to define additional parameters that may be present in the XML input.
The name at each level is concatenated with prior levels to form a unique instance identifier. The packet signature instances would thus be known as “Ethernet0.Kazaa.sig” and “Ethernet0.Foo.sig”. Individual data items from this provider will apply to the most granular name (e.g., Ethernet0.Foo.sig as opposed to simply Ethernet0).
Returning to the illustrative scenario, this data set defines the counters necessary to track the number of packets being generated for each signature on individual network adapters. The initial deployment to the customers might simply define the Kazaa signature. At a later point in time, the BSP may decide that they need to track Foo and provision a new file that defines a signature file and collection definition for that case.
In the second instance, no additional information beyond the network adapter name(s) are required.
Action modules 631 are modules that can be invoked in response to some trigger condition being activated.
The sample discussed herein leverages Microsoft NET namespaces and assemblies to define methods that can be called.
The action handler 627 reads the action definitions 609 and is responsible for managing the individual action providers 631.
Consider the sample input file.
The “id” attribute is simply a name for the action.
The “class” attribute defines a namespace and a class that can be created to invoke the action.
The “method” attribute defines a method (or function) within the class to invoke the action.
This concept could accommodate other types of action definitions, such as a COM creatable class with a set of IDispatch methods.
There may be additional “built in” actions provided by the action engine 635 itself that might, for example, upload the event to a specified upload target.
Metric modules 623 define aggregations of raw data collected in data sets. The metric modules 623 are specific data handlers that operate on the raw data to combine or analyze it in some way. The attributes on the primary metric element define how the metric is referenced later. The sub-elements define how the raw data is handled, what data fields are acted upon and from what providers they come.
Returning to our sample we see how we can combine multiple metrics from the tables and instances we have thus far defined to form a single combinational metric that we can use later in rules.
The sample defines a metric with the id PacketMonitor that is the running average of an equation that combines the data defined thus far. Window is a value specific to the RunningAverage metric type that defines how many samples the metric should be measured over.
Ultimately, the metrics defined here, while potentially rather complex, provide a simple result to rules that can then be able to perform simple comparisons against a threshold value and invoke an action in response.
There may be a set of metric types built into the tool 301. This set may include the aforementioned “RunningAverage” metric type. The tool 301 supports dynamic metric type providers in much the same way as actions are supported. This allows the tool 301 to be extended on the fly with new metric types.
The long set of XML beginning with <DivideStatement> defines a number that can be fed to running average and then to a rule.
The <DivideStatement> element has two children; one is a “point value” and the other is yet another arithmetic operation. This means that the first value is divided by the second to yield the final value that is fed to a running average component of the compositional metric collector 615.
Point Value refers back to a specific instance of a value thus far defined. It should be apparent that this refers back to the PacketCount row from the provider that analyzes packet signatures.
The <SumStatement> element is similar. It references two more points values pulled out of the network stats table. PktsOut and Pktsin added together result in the total number of packets through an adapter.
Therefore the result of the equation is the total number of packets matching a signature divided by the total number of packets through the network adapter (i.e., the percentage of packets that match a signature).
The PacketMonitor metric is thus the running average over 20 samples of this equation.
What may be unclear at this point is to what these metrics apply (i.e., which instances). Recall that there are two instances tracked by the packet signature provider: Ethernet0.Kazaa.sig and Ethernet0.Foo.sig. There is one instance tracked by the network provider: Ethernet0.
If all individual data items in the metric resolve to exactly the same instances, then it is fairly evident how many individual metrics will be generated on each collection of the metric.
If they are different, as in this case, the number of individual metrics will be equal to the number of metric instances produced by the data set with the greatest number of instances. In an embodiment, data item names must share the same prefix or they cannot be combined. To further clarify: We have a total number of packets through the adapter Ethernet0. We then have two counters relative to Ethernet0: Kazaa.sig and Foo.sig. Ultimately the two equations that we evaluate are:
Ethernet0.Kazaa.sig.PacketCount/(Ethernet0.PktsOut+Ethernet0.PktsIn)
Ethernet0.Foo.sig.PacketCount/(Ethernet0.PktsOut+Ethernet0.Pktsin)
Returning to the exemplary scenario, the BSP has now easily defined a method for tracking the percentage of packets on each adapter 105 that match a particular signature. The monitoring system 621 handles many of the details, and the BSP has largely produced a few simple components and a modified input file.
Furthermore, an additional metric identified as “NetworkAdapterUsage” is defined that measures the running average of network traffic through the adapter.
Rules define analyses on metrics that may result in an action. Information about when the rule fires, how it should be referenced elsewhere in the tool 301 and its description are among the attributes defined.
The monitoring system 621 is responsible for reading the definition of rule providers 625 and managing the flow of data coming from one or more of the metric providers 613, 623. Positive responses from rules result in the monitoring system 621 invoking one or more actions through the action handler 627.
Sub-elements define what metric(s) are used and to what they are compared to result in a triggering of the rule.
Consider the following sample:
This rule takes the PacketMonitor metric defined previously and compares it against the value 15. The “type” attribute indicates that it is a “greater than” threshold, so if the value of PacketMonitor rises above 15 the “PacketSigExceeded” action will be triggered.
Any of these actions may invoke the target manager 629 to forward data around the network 101 or to an external target (e.g., server 407) through the Internet 107.
The original scenario presented the problem wherein a teenage child in the family installed a P2P file sharing program that flooded the network with traffic and impacted mom's ability to engage in video chat sessions with Grandma.
There really was no information provided to the user or the BSP that would help them isolate why, after the router was installed, Mom suddenly experienced degraded performance.
An embodiment of the system described herein allows the BSP to provide a monitoring configuration file, such as the sample file described above herein, in response to the establishment of the home LAN and changing conditions in the Internet that will enable the BSP to more accurately respond to customer inquiries.
For example, if Mom calls the BSP in this scenario and this system has been running, it will be obvious that a significant portion of network traffic is dominated by the daughter's pc running this P2P application. The customer service representative can then instruct the customer to shut this application down on their teenager's computer and perhaps instruct them how to prevent their daughter from running as an administrator so she can't install such applications at all.
Of course, there's no reason why the system cannot be more proactive. Recall that rules trigger arbitrary actions that are defined in input files. The system could notify Mom's computer that this is happening and instruct her on how to remedy the situation without the involvement of the CSR.
An embodiment of the invention allows for metrics to be collected conditionally. This functionality allows, for example, a rule to trigger the collection of metrics for the purpose of diagnostics.
The monitoring system 621 can load new rules on the fly. For example, either a primary declaration file can be rebuilt and loaded or extra configuration files 601 can be loaded in on the fly.
An embodiment of the invention allows for aggregated data to be fed from the monitoring system 621 up to a higher level entity outside the home network 101, such as the BSP, for example. This allows the BSP to define data that is interesting, have it collected by the tool 301, and fed into a central repository, such as server 407. The BSP can analyze this data to develop new rules and metrics that can allow them to feed new schemas down to respective tools 301 running on customer networks that detect and solve problems programmatically. The target manager 629 facilitates communication with the network service provider/BSP.
Two other rules are defined:
These rules monitor how busy network adapters are both on a local device 103 and on other devices in the network 101. A new concept introduced in these rules is the “source” attribute. On the first rule “avgBandwidthRising,” the source is specified as local. This means that as raw metric tables are collected and shared around the network 101, this rule will only fire for raw data tables collected on this device 103. The other rule specifies the source as “remote” meaning that the rule will only fire for raw data tables collected from other devices on the network 101. The rule will fire locally but it will only be in response to receiving a table of raw data from another device 103 and evaluating the rule locally on that data.
State machines 626 define interactions of rules into a flowchart of states, any one of which might invoke an action. Consider the scenario where the user is notified if he is consuming a significant amount of bandwidth and someone else on the network 101 is as well. The following state machine definition describes how this situation might be detected:
Recall that if only one device 103 is busy, that is not a condition that requires notification to the user. However, if multiple devices 103 are busy, then the user may be impacted and an embodiment should detect this.
This example shows one StateMachine element with multiple states. In practice there could be multiple StateMachine elements. The first state, named “start” in the XML has rules that trigger a move to a new state. If the previously defined “avgBandwidthRising” rule is triggered, the state machine moves to the “ImBusy” state. This means that an embodiment has detected that the local computer is busy.
If the “avgOtherBandwidthRising” state is triggered, an embodiment moves to a state, “OtherBusy” where the tool 301 has detected that another computer on the network 101 is busy.
From these two states, the tool 301 can move to a state, “NetworkBusy” where it has detected that the local device 103 is busy and another device of the network 101 is busy thus requiring user notification. The “NetworkBusy” state has an attribute action that will be invoked when the state is entered. The “NotifyUser” action can be defined as previous actions are defined and might upload the fact that the state has been reached to the server 407 or might let the user know that a condition has arisen in which they might experience degraded performance.
In practice, triggers can also be defined that result in a return to a prior state. For example, if the state machine 626 is in the ImBusy state, then, if the bandwidth dropped back below the threshold, the state machine can transition back to the start state.
An embodiment of the invention may allow for defining how data can be shared with other monitoring system 621 instances around the network 101. The target manager 629 facilitates this sharing as well as uploading data across the Internet 107.
While embodiments of the invention have been described with respect to specific examples including presently preferred modes of carrying out the invention, those skilled in the art will appreciate that there are numerous variations and permutations of the above described systems and techniques that fall within the spirit and scope of the invention as described herein.
This application claims priority from U.S. Provisional Patent Application No. 60/949,628, filed Jul. 13, 2007, entitled “METHOD AND SYSTEM FOR MONITORING A HOME NETWORK,” which is hereby incorporated by reference in its entirety as if fully set forth herein.
Number | Date | Country | |
---|---|---|---|
60949628 | Jul 2007 | US |