Distributed system to determine a server's health

Abstract

Exemplary embodiments for a distributed system for determining a server's health are disclosed. The systems and methods provide for a network controller to direct one or more servicing nodes to check the health of one or more servers, and report a health score to the network controller. The network controller may then calculate, update and maintain a health score for each server in the network from the various health scores reported to it from the servicing nodes. This allows a distributed system to be used to facilitate network operations, as a single device is not relied on for periodically determining each server's health.

Description

BACKGROUND OF THE INVENTION

Field of the Invention

The present invention relates generally to data networks, and more particularly to components of a data network to determine a health score of a server.

Description of the Related Art

The tremendous growth of networked clients such as smartphones and tablets has led to a large growth of network services. Many network services require deployment of a substantial number of servers and server load balancers to handle the service load. It is not uncommon to see thousands of servers used for a single popular application such as a mobile messaging application, a mobile photo application, an Internet radio or music streaming services, or a multi-person online game. In such service deployment situations, the load put on the servers is not necessarily even and is often a function of service policies of connected server load balancers (SLB), traffic managers, or application delivery controllers (ADC). When a server becomes busy, it is desirable for an ADC to detect the situation and select a more available server for new service requests. In addition, servers may fail or be taken off the service by a network administrator for maintenance reasons. It is important for an ADC to be aware of such situations and to select other servers to handle the traffic.

Typically, an ADC is connected to all servers and often detects the health of all servers directly. Upon knowing the servers' health after the detection, the SLB or ADC would adjust their service application processing accordingly. Such direct detection mechanism works well when there are few servers. However, when there are thousands of servers, the computing and resource effort to check the servers' health will consume a significant part of available resources of the ADC. Since a typical service network may have another ADC, it would be advantageous to share the load of detecting the servers' health. An ADC can be assigned to check the health of a portion of the servers while it relies on other ADCs to determine the health of other servers.

It should be apparent from the foregoing that there is a need to provide a distributed method to determine the health of a server.

SUMMARY

This summary is provided to introduce a selection of concepts in a simplified form that are further described in the Detailed Description below. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

The present disclosure is related to approaches for a distributed system to determine a server's health. A method for determining, by a network controller, a health score for a server may comprise receiving a request for a service session from a server, directing at least one servicing node to determine a health score of a server, receiving a health score of the server from each of the at least one servicing nodes, calculating an aggregated health score using the received health scores, and sending the aggregated health score to a servicing node designated to process the service session. The servicing node may use the aggregated health score to select or not select the server for a service session. The servicing node that receives the aggregated health score may be a different servicing node than one that reported a health score for the selected server.

A distributed system for determining a health score for a server is also disclosed. The system may comprise a plurality of servicing nodes, and a network controller that is configured to receive a request for a service session from a server, direct at least one servicing node to determine a health score of a server, receive a health score of the server from each of the at least one servicing nodes, calculate an aggregated health score using the received health scores, and send the aggregated health score to a servicing node designated to process the service session.

In further example embodiments of the present disclosure, the method steps are stored on a machine-readable medium comprising instructions, which when implemented by one or more processors, perform the recited steps. In further example embodiments, hardware systems, or devices, can be adapted to perform the recited steps. Other features, examples, and embodiments are described below.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments are illustrated by way of example and not by limitation in the figures of the accompanying drawings, in which like references indicate similar elements.

FIG. 1 illustrates an exemplary embodiment of a data network servicing a service session based on server health.

FIG. 2 illustrates an exemplary embodiment of a network node.

FIG. 3 illustrates an exemplary embodiment of a network controller collecting a server's health scores.

FIG. 4 illustrates an exemplary embodiment of a server health score.

FIG. 5 illustrates an exemplary embodiment of a network controller calculating an aggregated health score for a server.

FIG. 6 illustrates an exemplary embodiment of a servicing node using a health score to process a service session.

DETAILED DESCRIPTION

The following detailed description includes references to the accompanying drawings, which form a part of the detailed description. The drawings show illustrations in accordance with example embodiments. These example embodiments, which are also referred to herein as “examples,” are described in enough detail to enable those skilled in the art to practice the present subject matter. The embodiments can be combined, other embodiments can be utilized, or structural, logical, and electrical changes can be made without departing from the scope of what is claimed. The following detailed description is therefore not to be taken in a limiting sense, and the scope is defined by the appended claims and their equivalents.

FIG. 1 illustrates an exemplary embodiment of a data network handling a service session using a service based on server health. Client 105 conducts a service session 110 with a server, such as server 115, server 120, or server 125. Although three servers are depicted in FIG. 1, there can be any number of servers. In one embodiment, data packets of service session 110 are processed by service network 130 prior to being delivered to server 115. Service network 130 comprises one or more servicing nodes, such as servicing nodes 135, 140, 145, and 150. Additionally, service network 130 connects to one or more servers 115, 120, and 125. In some embodiments, service network 130 connects to a network controller 155. In various embodiments, network controller 155 may also perform functions of a servicing node.

The one or more servicing nodes may examine the health of the servers. In an exemplary embodiment, servicing node 150 is assigned to check the health of server 115 and report a health score for server 115 to network controller 155. In some embodiments, servicing node 150 is not assigned to examine server 115 and does not report the health score of server 115 to network controller 155. Network controller 155 collects reported health scores of server 115 from the one or more servicing nodes, calculates server health score 160 for server 115, and stores server health score 160. Network controller 155 also collects reported health scores of servers 120 and 125, and calculates server health scores for those servers as well.

In various embodiments, network controller 155 sends server health score 160 to a servicing node, such as servicing node 135. Servicing node 135 determines if server 115 is healthy according to server health score 160. Servicing node 135 processes service session 110 and selects server 115 to process service session 110 if server 115 is determined to be healthy. Servicing node 135 then sends service session 110 to server 115 for processing.

Alternatively, network controller 155 may send server health score 160 for server 115 to a different servicing node, such as servicing node 150. Servicing node 150 determines server health score 160 to be poor and thus server 115 not healthy according to server health score 160. Servicing node 150 processes service session 110 and does not select server 115 based on the server health score 160. Instead, servicing node 150 selects a different server, such as server 120, to handle the service session 110.

In various embodiments, service network 130 includes additional network elements such as routers, switches, Ethernet switches, ATM switching, wireless network elements, broadband gateways, top of rack switches, access points, base stations, mobile network elements, data traffic controller, application delivery controllers, firewalls, optical switches, remote access gateways and other data network gateways, or any other network elements. Additionally, service network 130 may utilize network protocols such as IP protocol, TCP/UDP, ATM, Frame Relay, Ethernet, VLAN, WLAN, MPLS, HTTP, and other data communication protocols.

FIG. 2 illustrates an exemplary embodiment of a network node 205 in a service network. In various embodiments, network node 205 is a client device, a server device, a servicing node, or a network controller. Network node 205 may include a processor module 210, a network module 215, and a computer storage module 220. Processor module 210 can include one or more processors, including a micro-processor, an Intel processor, an AMD processor, a MIPS processor, an ARM-based processor, a RISC processor, or any other type of processor. Processor module 210 can also include one or more processor cores embedded in a processor. Additionally, processor module 210 can include one or more embedded processors, or embedded processing elements in a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), or Digital Signal Processor (DSP).

In exemplary embodiments, network module 215 comprises a network interface such as Ethernet, optical network interface, a wireless network interface, T1/T3 interface, or a WAN or LAN interface. Furthermore, network module 215 can include a network processor. Computer storage module 220 comprises RAM, DRAM, SRAM, SDRAM, or any other memory utilized by processor module 210 or network module 215. Computer storage module 220 stores data utilized by processor module 210. Computer storage module 220 can include a hard disk drive, a solid state drive, an external disk, a DVD, a CD, a readable external disk, or any other type of storage. Additionally, computer storage module 220 stores one or more computer programming instructions, which when executed by processor module 210 or network module 215, implements one or more of the functionalities of the present invention. Network node 205 can also include an input/output (I/O) module 225, such as a keyboard, a keypad, a mouse, a gesture-based input sensor, a microphone, a physical or sensory input peripheral, a display, a speaker, a physical or sensory output peripheral, or any other input or output module.

Network node 205 may also include service application 230. When network node 205 is a client, it performs a client-side function of service application 230. A client-side function of service application 230 comprises file downloading, file uploading, web browsing, video streaming, e-book reading, game playing, gaming, photo applications, messaging, voice calling, music listening, conferencing, document editing, map navigation, online shopping, and other client-networking activities. On the other hand, service application 230 may include server-side functionalities when network node 205 functions as a server. Server-side functions of service application 230 comprise file transfer, HTTP server functions, voice call services, video and audio streaming, online games, collaboration, enterprise work flow, enterprise application, enterprise resource planning (ERP), customer relationship management (CRM), conferencing, e-commerce, transaction processing, business to business (B2B) applications, machine to machine (M2M) application, telemetry, remote access, identity authentication, and other server application functionalities.

In various embodiments, service application 230, when executed in network node 205 as a servicing node, comprise functions such as a network proxy application, security gateway, firewall, protocol gateway application, server load balancing, content caching, network optimization, data de-duplication, network cached storage access, browser optimization, and other network-based service applications. Service application 230 of a servicing node may also include functionalities implementing a part of the present invention.

Service application 230, when executed in network node 205 as a network controller, comprises network management functionalities over service network 130 such as provisioning, network node configuration, event logging, network center operation, network statistics collection and/or analysis, alert handling, and/or other network operation and management applications.

FIG. 3 further illustrates an exemplary embodiment of network controller 155 working in conjunction with one or more servicing nodes to determine a server health score. The one or more servicing nodes may include servicing nodes 135, 140, 145, and 150 and are connected to one or more servers, such as server 115.

In some embodiments, network controller 155 informs servicing nodes 135, 140, and 145 to report the health score of server 115. In other embodiments, network controller 155 informs servicing node 150 to not report the health score of server 115, or servicing node 150 does not receive any instruction from network controller 155 to report the health score of server 115, and therefore servicing node 150 does not report the health score of server 115. In various embodiments, servicing node 135, similarly for servicing node 140 or 145, checks the health of server 115 by querying server 115 and then generates server health score 305 for server 115.

Server health score 305, as illustrated in an exemplary embodiment in FIG. 4, comprises a score for connectivity or reachability, resource utilization such as CPU and memory usage, load of a service application, or various measures for the load of the service application. The various measures for the load of the service application comprise the number of active sessions or connections, pending sessions, or completed sessions; performance of the service application such as page loading time; or average, maximum, or minimum response time. Server health score 305 may also include one or more internal measurements within a servicing node, related to the server. Such internal measurements comprise data traffic sent from the servicing node to the server, data traffic from a client device using the servicing node directed to the server, or data traffic processed by the servicing node for a service application of the server.

In various embodiments, a servicing node uses different means to determine the measures indicated in server health score 305. For example, servicing node 135 may connect to server 115 using a network link. Servicing node 135 checks or measures the health of the network link to determine connectivity indicated in server health score 305. Servicing node 135 may also connect to server 115 over service network 130. Servicing node 135 checks reachability to server 115 by sending a query using a protocol, such as ping packet, echo or a proprietary protocol, to server 115 and examining if server 115 responds to the query. Servicing node 135 measures and records the reachability status into server health score 305.

In some embodiments, servicing node 135 sends a request to server 115 to obtain resource utilization information. Servicing node 135 sends a piece of software or computer programming instructions to server 115 which executes the sent software to collect and report system resource utilization to servicing node 135. In various embodiments, to obtain a load of the service application, servicing node 135 sends a service request, such as a HTTP request, for the service application and calculates the service response from server 115 with a pre-stored response. Servicing node 135 compares the performance of the service request with a pre-stored expectation of the performance. Server 115 sends to servicing node 135 a measure about the working load of the service application in a response to a client service request, separately, or in a HTTP response to servicing node 135.

In some embodiments, a servicing node 135 sends all the information in server health score 305 to network controller 155. Servicing node 135 calculates a summary server health score 305, such as “Good”, “Bad”, “Poor”, “Failed”, “1,” or “0,” and sends server health score 305 to network controller 155.

In an exemplary embodiment illustrated in FIG. 5, network controller 155 maintains server health score 160 which indicates health of server 115. Server health score 160 stores a list of health scores reported by one or more servicing nodes such as servicing nodes 135, 140, 145, or 150.

In an exemplary embodiment, network controller 155 receives server health score 305 reported from servicing node 135 about health of server 115. Network controller 155 stores server health score 305 into the corresponding entry in server health score 160. Entry for servicing node 150 is not applicable when network controller 155 determines servicing node 150 does not need to report health score for server 115.

Network controller 155 determines if there is sufficient information in server health score 160, and network controller 155 calculates aggregated health score 505 for server health score 160. In various embodiments, network controller 155 determines there is sufficient information in server health score 160 when it receives a server health score 305 from at least one servicing node, from a majority of servicing nodes, from at least one servicing node within a specified time frame (e.g. last 30 seconds, last 5 seconds), from at least two servicing nodes within a specified time frame, from all servicing nodes, or any similar variation. Aggregated health score 505 may indicate combined health information about server 115. Network controller 155 sends aggregated health score 505 for server 115 to servicing nodes 135, 140, 145, or 150.

Network controller 155 may calculate aggregated health score 505 by computing an average, weighted average, mean, median, or other arithmetic operation of the received server health scores from the at least one servicing node. As such, the aggregate health score 505 may be different from the server health score 160. In some embodiments, aggregate health score 505 calculated by the network controller 155 is also included in server health score 160. In various embodiments, network controller 155 determines aggregated health score 505 to be good, “1”, or “yes” when a majority of received server health scores indicate good health for server 115. On the other hand, network controller 155 determines aggregated health score 505 to be poor, “0”, or “no” when a majority of received server health scores indicate poor health for server 115. Network controller 155 considers different weighing factors to different reported server health scores depending on the reporting servicing node. In some embodiments, the calculation is based on unanimous agreement of reported server health scores. In other embodiments, network controller 155 discards a reported health score prior to the calculation, or when a reported health score is not trusted.

In an exemplary embodiment, network controller 155 receives server health score 305, replaces an earlier reported health score from servicing node 135, and calculates an updated aggregated health score 505. Network controller 155 calculates an updated aggregated health score 505 when network controller 155 receives a sufficient number of newly received server health scores. Network controller 155 may include a timer (not shown) and discards all current reported health scores. After receiving another collection of reported server health scores from the one or more servicing nodes, network controller 155 may calculate an updated aggregated health score 505.

In some embodiments, network controller 155 stores a history of reported server health scores. Network controller 155 determines from a history of reported server health scores from servicing node 135 whether the reported server health scores are not consistent and whether servicing node 135 is not trusted for further reporting server health score for a particular server, such as server 115. Then, network controller 155 instructs servicing node 135 not to further report server health score, and network controller 155 discards reported health scores from servicing node 135 in calculating aggregated health score 505. Additionally, network controller 155 instructs a different servicing node, such as servicing node 150, to report server health for server 115, so as to replace the role of servicing node 135.

In some embodiments, network controller 155 determines servicing node 135 is no longer needed to report server health for server 115, and network controller 155 instructs servicing node 135 not to report server health for server 115. Servicing node 135 may inform network controller 155 that it no longer will report server health for server 115. Upon determining that servicing node 135 is no longer needed to report server health for server 115, network controller 155 removes an entry in server health score 160 corresponding to servicing node 135.

In an exemplary embodiment illustrated in FIG. 6, servicing node 135 determines a final server health score 605 for server 115. Servicing node 135 determines server health score 305 for server 115, reports server health score 305 to network controller 155, or receives server health score 610 from network controller 155. In various embodiments, server health score 610 is the same as server health score 160, or aggregate health score 505. Servicing node 135 combines received server health score 610 and server health score 305 to get final server health score 605.

In some embodiments, servicing node 135 may receive server health score 610 from network controller 155, and detect server health score 305 directly from the server. The servicing node makes its own determination as to which server health score to use as final server health score 605. In an exemplary embodiment, network controller 155 receives a “good” server health score from servicing nodes 135 and 140, and a “bad” server health score from servicing node 145. Network controller 155 determines aggregate health score 505 to be “good”, since the majority of the received server health score are good. Servicing node 145 receives aggregate health score 505 of “good”, which conflicts with its own detected value of “bad”, and then makes its own determination of which value to use as final server health score 605. Conflicting server health scores for server 115 may be due to the servicing nodes checking the health of server 115 at different times, using different means. Network controller 155 also calculates and updates aggregated health score 505 using scores received at different times from different servicing nodes.

In some embodiments, servicing node 135 does not use server health score 610 received from network controller 155 and considers only server health score 305 to determine final server health score 605. In other embodiments, server health score 610 is more accurate, and servicing node 135 considers only server health score 610. If server health score 610 is much worse than server health score 305, conflicts with server health score 305, or indicates that server 115 is not available or failed, servicing node 135 may use only server health score 610. In various embodiments, servicing node 135 is not assigned to report health score for server 115, and servicing node 135 uses only server health score 610 to determine final server health score 605.

In some embodiments, servicing node 135 processes service session 110 using final server health score 605 and a service policy 615. According to service policy 615, servicing node 135 selects server 115 to process service session 110. Servicing node 135 checks final server health score 605 of server 115. If final server health score 605 indicates server 115 is available, servicing node 135 selects server 115 to process service session 110. If final server health score 605 indicates server 115 is not available, busy, congested, or failed, servicing node 135 does not select server 115. In various embodiments, health server score 305 may indicate that server 115 is available such that servicing node 135 would have selected server 115, while final server health score 605 indicates server 115 is not available. In those embodiments, servicing node 135 does not select server 115 to process service session 110. In other embodiments, servicing node 135 does not detect a health score for server 115 such that servicing node 135 would have selected server 115 while final server health score 605 indicates server 115 is not available. In those embodiments, servicing node 135 does not select server 115 to process service session 110.

Servicing node 135 examines service policy 615 and select another server such as server 120 to process service session 110. In some embodiments, servicing node 135 removes server 115 from service policy 615 based on final server health score 605 so that servicing node 135 would not consider server 115 when processing service session 110. Alternatively, servicing node 135 may add server 115 to service policy 615 based on final server health score 605 so that servicing node 135 would consider server 115 when processing service session 110.

In various embodiments, servicing node 135 receives an updated server health score 610 from network controller 155 and calculates an updated final server health score 605. In other embodiments, servicing node 135 determines server health score 305 and calculate an updated final server health score 605. Servicing node 135 updates service policy 615 based on an updated final server health score 605.

In some embodiments, network controller 155 assigns different servicing nodes to collect server health scores for different servers. Network controller 155 determines that server 115 is more important than server 120 and assigns more servicing nodes for server 115 than for server 120. If network controller 155 observes that aggregated health scores of server 115 are inconsistent over a period of time, network controller 155 may assign additional servicing nodes to collect health scores of server 115, so as to better determine the aggregated health of server 115. However, if network controller 155 observes that aggregated health scores of server 115 are stable over a period of time, network controller 155 may reduce a number of servicing nodes to monitor server 115, so as to reduce the processing load of the servicing nodes, the server, and the network. In various embodiments, network controller 155 increases the frequency of health data collection by the servicing node. Alternatively, network controller 155 may decrease the frequency of health data collection, depending on the trend of aggregated health of server 115 over a period of time.

In some embodiments, network controller 155 connects to a display and displays the aggregated health score of server 115 and the trend of aggregated health of server 115. Network controller 155 generates a report of the history of aggregated health scores of one or more servers. Network controller 155 applies one or more rules to analyze the history of aggregated health scores to determine a trend of server loads, a trend of service application load, or a trend of client traffic.

Claims

1. A method for determining, at a network controller, a health score of a server, the method comprising: receiving, from a client, a request for a service session with a server of a plurality of servers, each of the plurality of servers being in communication with one or more servicing nodes of a plurality of servicing nodes;directing the plurality of servicing nodes in communication with the network controller to determine a health score of at least one of the plurality of servers, each of the plurality of servicing nodes being configured to determine the health score of the one or more of the plurality of servers, the determining of the health score by each of the plurality of servicing nodes including determining a server health score of the one or more of the plurality of servers based at least on reachability of the one or more of the plurality of servers, the reachability being determined by sending a query to the one or more of the plurality of servers and examining whether the one or more of the plurality of servers respond to the query;receiving the health score of the at least one server from each of the plurality of servicing nodes;calculating an aggregated health score for each server of the plurality of servers using the received health scores from the plurality of servicing nodes; andsending the aggregated health score to a servicing node of the plurality of servicing nodes, the servicing node being designated to process the service session by selecting, based at least on the aggregated health score, one of the plurality of servers for providing the service associated with the service session and directing the request for the service session to the selected one of the plurality of servers, wherein the servicing node designated to process the service session is not one of the plurality of servicing nodes that reported the health score of the one of the plurality of servers to the network controller.

2. The method of claim 1, wherein the sending the aggregated health score to the servicing node designated to process the service session enables the servicing node to use the aggregated health score to select the server or not to select the server for the service session.

3. The method of claim 2, wherein the servicing node designated to process the service session is one of the plurality of servicing nodes directed to determine a health score of at least one of the plurality of servers, and the aggregated health score conflicts with the health score determined by the servicing node.

4. The method of claim 3, wherein the servicing node designated to process the service session does not select the server based on the aggregated health score when the servicing node designated to process the service session would have selected the server if using the health score determined by the servicing node.

5. The method of claim 1, wherein, based on the aggregated health score, the servicing node designated to process the service session does not select the server, when the servicing node would have selected the server without using the aggregated health score.

6. The method of claim 1, wherein the aggregated health score is based on a majority of health scores received by the network controller.

7. The method of claim 6, wherein the majority of received health scores indicate that the server is in good health, or the majority of received health scores indicate that the server is in poor health.

8. The method of claim 1, wherein the calculating is based on an average of the received health scores, or a weighted average of the received health scores.

9. The method of claim 1, further comprising: receiving one or more updated health scores from at least one of the plurality of servicing nodes;calculating an updated aggregated health score based on the one or more updated health scores; andsending the updated aggregated health score to the servicing node designated to process the service session.

10. The method of claim 1, wherein at least one of the received health scores comprises one or more scores related to connectivity, system resource utilization, number of connections, and load of a service application.

11. A non-transitory computer-readable storage medium comprising instructions, which when executed by one or more processors of a network controller, perform a method for determining a health score of a server, the method comprising: receiving, from a client, a request for a service session with a server of a plurality of servers, each of the plurality of servers being in communication with one or more servicing nodes of a plurality of servicing nodes;directing the plurality of servicing nodes in communication with the network controller to determine a health score of at least one of the plurality of servers, each of the plurality of servicing nodes being configured to determine the health score of the one or more of the plurality of servers, the determining of the health score by each of the plurality of servicing nodes including determining a server health score of the one or more of the plurality of servers based at least on reachability of the one or more of the plurality of servers, the reachability being determined by sending a query to the one or more of the plurality of servers and examining whether the one or more of the plurality of servers respond to the query;receiving the health score of the at least one server from each of the plurality of servicing nodes;calculating an aggregated health score for each server of the plurality of servers using the received health scores from the plurality of servicing nodes; andsending the aggregated health score to a servicing node of the plurality of servicing nodes, the servicing node being designated to process the service session by selecting, based at least on the aggregated health score, one of the plurality of servers for providing the service associated with the service session and directing the request for the service session to the selected one of the plurality of servers, wherein the servicing node designated to process the service session is not one of the plurality of servicing nodes that reported the health score of the one of the plurality of servers to the network controller.

12. A system for determining a health score of a server, the system comprising: a plurality of servicing nodes in communication with a plurality of servers that provide a service; anda network controller comprising a processor and a memory, the network controller in communication with the plurality of servicing nodes, the network controller configured to:receive, from a client a request for a service session from a client with a server of a plurality of servers, each of the plurality of servers being in communication with one or more servicing nodes of the plurality of servicing nodes:direct at least one the plurality of servicing nodes to determine a health score of at least one of the plurality of servers, each of the plurality of servicing nodes being configured to determine the health score of the one or more of the plurality of servers, the determining of the health score by each of the plurality of servicing nodes including determining a server health score of the oneor more of the plurality of servers based at least on reachability of the one or more of the plurality of servers, the reachability being determined by sending a query to the one or more of the plurality of servers and examining whether the one or more of the plurality of servers respond to the query;receive the health score of the at least one server from each of the at least one plurality of servicing nodes;calculate an aggregated health score for each server of the plurality of servers using the received health scores; andsend the aggregated health score to a servicing node of the plurality of servicing nodes, the servicing node being designated the process the service session by selecting, based at least on the aggregated health score, one of the plurality of servers for providing the service associated with the service session and directing the request for the service session to the selected one of the plurality of servers, wherein the servicing node designated to process the service session is not one of the plurality of servicing nodes that reported the health score of the one of the plurality of servers to the network controller.

13. The system of claim 12, wherein the network controller configured to send the aggregated health score to the servicing node designated to process the service session enables the servicing node to use the aggregated health score to select the server or not select the server for the service session.

14. The system of claim 12, wherein the servicing node designated to process the service session is one of the at least one servicing nodes, and the aggregated health score conflicts with the health score determined by the servicing node.

15. The system of claim 12, wherein the aggregated health score is based on a majority of health scores received by the network controller.

16. The system of claim 12, wherein the aggregated health score is based on an average of the received health scores, or a weighted average of the received health scores.

17. The system of claim 12, wherein the network controller is further configured to: receive one or more updated health scores from the at least one servicing nodes;calculate an updated aggregated health score based on the one or more updated health scores; andsend the updated aggregated health score to the servicing node designated to process the service session.

18. The system of claim 12, wherein at least one of the received health scores comprises one or more scores related to connectivity, system resource utilization, number of connections, and load of a service application.