System for distributing load over multiple servers at an internet site

• US 6,317,775 B1
• Filed: 01/25/1999
• Issued: 11/13/2001
• Est. Priority Date: 11/03/1995
• Status: Expired due to Term

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Tenereillo, Peter A., Kersey, Edward C., Coile, Brantley W., Howes, Richard A.
• Primary Examiner(s): Follansbee, John A.
• Application Number:
  US08/850,248
• Time in Patent Office: 1,023 Days
• Field of Search: 709/201, 709/226, 709/217, 709/223, 702/182, 370/238
• US Class Current: 709/201
• CPC Class Codes: H01R 2201/04 for network, e.g. LAN connectors
• CPC Class Codes: H01R 31/005 Intermediate parts for distributing signals
• CPC Class Codes: H04L 2101/663 Transport layer addresses, e.g. aspects of transmission control protocol [TCP] or user datagram protocol [UDP] ports
• CPC Class Codes: H04L 41/00 Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
• CPC Class Codes: H04L 61/00 Network arrangements, protocols or services for addressing or naming
• CPC Class Codes: H04L 61/2514 between local and global IP addresses
• CPC Class Codes: H04L 61/2564 for a higher-layer protocol, e.g. for session initiation protocol [SIP]
• CPC Class Codes: H04L 63/0254 Stateful filtering
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• 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 67/1017 based on a round robin mechanism
• CPC Class Codes: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

Disclosed is a system and method for distributing connections among a plurality of servers at an Internet site. All connections are made to a single IP address and a local director selects the server from among the plurality of servers which is to receive the connection. Thus, the DNS server is not relied upon to distribute connections, and the connection distribution scheme is not avoided when DNS is bypassed. In one embodiment, a session distribution scheme is implemented such that connections are distributed to the server in the group of servers which has the fewest connections of the group. In other embodiments, other session distribution schemes which route connections based on the predicted response times of the servers or according to a round robin scheme are used.

Claims

1. A system for distributing connections from clients on an external network to a plurality of servers on an internal network, the system comprising:

a client interface to the external network the client interface being operative to receive and send packets to and from a remote client;

a server interface to the internal network, the server interface being operative to receive and send packets to and from a plurality of servers, the plurality of servers being operative to establish a connection with the remote client;

a plurality of predicted responsiveness indicators, each of the plurality of predicted response indicators being associated with each of the plurality of servers, the predicted responsiveness indicators being operative to predict the response time of each of the plurality of servers, the predicted responsiveness indicators also being stored within the system in a manner that the predicted responsiveness indicators may be accessed; and

a predicted responsiveness comparator which is operative to access and compare the predicted responsiveness indicators and to determine which servers from among the plurality of servers is associated with a predicted responsiveness indicator which measures a best response time, the predicted responsiveness comparator being further operative to select a pointer to a server which has a predicted responsiveness that is the best predicted responsiveness among the predicted responsiveness of the plurality of servers;

wherein the predicted responsiveness indicators include the predicted response time of each of the plurality of servers and wherein the predicted response time is calculated according to the formula;

2. A system as recited in claim 1, wherein the predicted responsiveness indicators are periodically updated.

3. A system as recited in claim 1, wherein the predicted responsiveness indicators include the number of connections to each of the plurality of servers.

4. A system as recited in claim 1, wherein the predicted responsiveness indicators include the number of connections to each of the plurality of servers.

5. A system as recited in claim 1, wherein the predicted responsiveness indicators include the predicted response time of each of the plurality of servers.

6. A system as recited in claim 1 wherein the aging function decreases the predicted response time as the time since the last predicted response increases.

7. A system as recited in claim 1 wherein R is a fimction of a plurality of measured response times.

8. A system as recited in claim 1 wherein R changes with each new measured response time, but is constrained to change by less than a certain percentage as a result of the latest measured response time.

9. A system as recited in claim 8 wherein the percentage is less than about 10%.

10. A system as recited in claim 8 wherein the percentage is less than about 25%.

11. A system as recited in claim 8 wherein the percentage is less than about 50%.