Client load distribution

• US 8,499,086 B2
• Filed: 01/21/2004
• Issued: 07/30/2013
• Est. Priority Date: 01/21/2003
• Status: Active Grant

• Current Assignee: Dell Products LP (Dell Technologies Inc.)
• Original Assignee: Dell Products LP (Dell Technologies Inc.)
• Inventors: Koning, G. Paul, Hayden, Peter C., Long, Paula, Suman, Daniel E.
• Primary Examiner(s): Etienne, Ario
• Assistant Examiner(s): NANO, SARGON N
• Application Number:
  US10/762,985
• Publication Number:
  US 20040215792A1
• Time in Patent Office: 3,478 Days
• Field of Search: 709/217, 709/223, 709/224, 709/225, 709238-242, 709/219, 709/229
• US Class Current: 709/229
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/10015 Access to distributed or replicated servers, e.g. using brokers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1029 using data related to the state of servers by a load balancer
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

Systems and methods for providing an efficient partitioned resource server. In one embodiment, the partitioned resource server comprises a plurality of individual servers, and the individual servers appear to be equivalent to a client. Each of the individual servers may include a routing table that includes a reference for each resource that is maintained on the partitioned resource server. Requests from a client are processed as a function of the routing table to route the request to the individual server that maintains or has control over the resource of interest.

Claims

1. A system for managing a set of connections between a plurality of clients and a plurality of storage servers based on system load, comprising:

a plurality of storage servers having a set of resources partitioned thereon, each of the storage servers in the plurality further comprising;

at least one connection of the set of connections being a client communication connection from the storage server to at least one client, the at least one connection further requesting a given resource of the set of resources;

a storage device for providing at least one volume of storage partitioned across the plurality of storage servers, such that a first portion of the given resource is located on a storage device associated with a first storage server of the plurality of storage servers, and such that a second portion of the given resource is located on a storage device associated with a second storage server of the plurality of storage servers;

a load monitor process connected to communicate with other load monitor processes on at least one other storage server of the plurality of storage servers for generating a measure of overall system load, and for generating a measure of client load on each of the plurality of storage servers that have at least one portion of the given resource; and

a client connection distribution process, responsive to the load monitor process, for repartitioning the set of connections and redistributing client load by disconnecting the at least one client communication connection from the first storage server and establishing a new connection from the at least one client communication connection to the second storage server based on a measurement of overall system efficiency determined from at least one of the measurement of overall system load, a category of the given resource requested, historical data available to the load monitor process in the plurality of storage servers relating to a previous access pattern from at least one storage server, policy parameters established for the plurality of servers, and the client load on each of the plurality of storage servers.

2. A system according to claim 1, further comprising:

a load distribution process for determining resource loads when establishing new connections among storage servers.

3. A system according to claim 1, further comprising:

a client allocation process for causing a client to communicate with a selected one of said plurality of servers.

4. A system according to claim 1, further comprising:

a client allocation process for distributing incoming client requests across said plurality of servers.

5. A system according to claim 1, wherein the client distribution process includes a round robin distribution process.

6. A system according to claim 1, wherein the client distribution process includes a disconnect process for dynamically disconnecting a client from a first server and reconnecting to a second server.

7. A system according to claim 1, wherein the client distribution process adaptively distributes new client connections across the plurality of servers as a function of dynamic variations in measured system load.

8. A system according to claim 1, further comprising:

a storage device for providing storage resources to the plurality of clients.

9. A storage area network, comprising:

a plurality of servers each configured as a server of claim 1.

10. An information storage server apparatus comprising:

one or more client communication connections to a plurality of client devices;

a server communication connection to at least one other storage server;

a storage device for storing a portion of a partitioned resource thereon, such that a first portion of a partitioned resource is stored on the storage server, and such that other portions of the partitioned resource are stored on the one other storage server;

a request handler, for receiving requests for access to the partitioned resource via at least one client communication connection, and determining which portions of the partitioned resource are stored on the storage device, and which portions of the partitioned resource are stored on the one other storage server;

a load monitor, for;

determining a local load placed on the storage server by the request handler;

receiving information via the server communication connection concerning a load on the other server; and

generating a measure of load based on at least one of a category of the given resource requested, historical data available to the load monitor process in the plurality of storage servers relating to a previous access pattern from at least one storage server, policy parameters established for the plurality of servers, and the client load on each of the plurality of storage servers from both the local load and the load on the other server; and

a client load distributor, responsive to the load monitor, for closing a client communication connection to at least one client device and requesting the at least one storage server to open a new client communication connection to the at least one client device when the measure of load indicates the storage server is over utilized and overall system efficiency may be improved.

11. The apparatus of claim 10 wherein the partitioned resource is at least one of a file, a file directory, a data block, a page, a database, or an application program.

12. The apparatus of claim 10 wherein the storage server and the other storage server are addressable by the client as a server group.

13. The apparatus of claim 12 wherein closing of the client communication connection by the client load distribution process is transparent to the client.

14. The apparatus of claim 10 additionally comprising:

a routing table, for storing information concerning which portions of the partitioned resource are stored on the storage server and which portions of the partitioned resource are stored on the other storage server.

15. The apparatus of claim 10 wherein the storage server and the other storage server are part of a storage area network.

16. The apparatus of claim 10 wherein the storage server and the other storage server are equivalent servers such that each equivalent server comprises a request handler, a load monitor and a client load distributor.

17. A method for handling information by providing a storage server comprising:

connecting to a plurality of client devices over one or more client communication connections;

connecting to at least one other storage server over a server communication connection;

storing a portion of a partitioned resource on a storage device local to the storage sever, such that a first portion of the partitioned resource is stored on the storage server, and such that at least one other portion of the partitioned resource is stored on the other storage server;

receiving requests for access to the partitioned resource from at least one of the clients, and determining if portions of the requested partitioned resource are stored on the storage device, and determining if portions of the requested portioned resource are stored on the other storage server and, if at least one portion of the requested resource is stored on either the storage device or the other server, directing the request to be served by a selected one of the storage server or the other storage server;

measuring a load by

determining a local load placed on the storage server based on the requests for access to the partitioned resource;

receiving information via the server communication connection concerning a load on the other server; and

generating a load measure based on at least one of a category of the given resource requested, historical data available to the load monitor process in the plurality of storage servers relating to a previous access pattern from at least one storage server, policy parameters established for the plurality of servers, and the client load on each of the plurality of storage servers using both the local load and the load on the other server; and

distributing client load by closing a client communication connection from one of the storage server or the other server to at least one client device, and requesting the other one of the storage server or the other server to open a new connection to the client device when the load measure indicates the storage server is over utilized and overall system efficiency may be improved.

18. The method of claim 17 wherein the partitioned resource is at least one of a file, a file directory, a data block, a page, a database, or an application program.

19. The method of claim 17 wherein the storage server and the other storage server are addressed as a server group by the client.

20. The method of claim 17 wherein closing the client communication connection by the client load distribution process is transparent to the client.

21. The method of claim 17 additionally comprising:

storing information in a routing table concerning which portions of the partitioned resource are stored on the storage server and which portions of the partitioned resource are stored on the other storage server.

22. The method of claim 17 wherein the storage server and the other storage server are part of a storage area network.

23. The method of claim 17 wherein the storage server and the other storage server are equivalent servers such that each perform the steps of request handling, measuring load and distributing client load.

24. A tangible, non-transitory, computer readable medium for storing computer executable instructions for providing a storage server function, with the computer executable instructions for:

connecting to a plurality of client devices over one or more client communication connections;

connecting to at least one other storage server over a server communication connection;

storing a portion of a partitioned resource on a storage device local to the storage sever, such that a first portion of the partitioned resource is stored on the storage server, and such that at least one other portion of the partitioned resource is stored on the other storage server;

receiving requests for access to the partitioned resource from at least one of the clients, and determining if portions of the requested partitioned resource are stored on the storage device, and determining if portions of the requested portioned resource are stored on the other storage server and, if at least one portion of the requested resource is stored on either the storage device or the other server, directing the request to be served by a selected one of the storage server or the other storage server;

measuring a load by

determining a local load placed on the storage server based on the requests for access to the partitioned resource;

receiving information via the server communication connection concerning a load on the other server; and

generating a load measure based on at least one of a category of the given resource requested, historical data available to the load monitor process in the plurality of storage servers relating to a previous access pattern from at least one storage server, policy parameters established for the plurality of servers, and the client load on each of the plurality of storage servers using both the local load and the load on the other server; and

distributing client load by closing a client communication connection from one of the storage server or the other server to at least one client device, and requesting the other one of the storage server or the other server to open a new connection to the client device when the load measure indicates the storage server is over utilized and overall system efficiency may be improved.

25. An information storage server apparatus comprising:

one or more network interfaces providing communication connections to a plurality of client devices and a server communication connection to a plurality of other storage servers;

a storage device for storing a portion of a partitioned resource thereon, such that a first portion of a partitioned resource is only stored on the storage server, and such that a second portion of the partitioned resource is only stored on one of the plurality of other servers;

opening an initial connection between one of the client devices and the storage server only if the storage server actually stores at least the first portion of the partitioned resource;

a request handler, for receiving requests for access to the partitioned resource via at least one client device, and determining if portions of the partitioned resource are stored on the storage device, and if portions of the partitioned resource are stored on the other storage server;

a load monitor, for;

determining a local load placed on the storage server by the request handler;

receiving information via the server communication connection concerning a load on the other server; and

generating a measure of load based on at least one of a category of the given resource requested, historical data available to the load monitor process in the plurality of storage servers relating to a previous access pattern from at least one storage server, policy parameters established for the plurality of servers, and the client load on each of the plurality of storage servers from both the local load and the load on the other server; and

a client connection distributor, responsive to both the request handler and the load monitor, for closing a client communication connection from the storage server to the client device and requesting that one of the plurality of other storage servers open a new client communication connection to the at client device when the measure of load indicates the storage server is over utilized.