Browse Prior Art Database

Redundancy Implementation using OSPF weighting

IP.com Disclosure Number: IPCOM000022653D
Original Publication Date: 2004-Mar-23
Included in the Prior Art Database: 2004-Mar-23

Publishing Venue

Motorola

Related People

Leonard K. Pennock: AUTHOR [+2]

Abstract

This paper outlines a method using OSPF to assign traffic to a favored server in a redundant server pool based on the states of the available servers and provides a method to avoid unwanted role reassignment.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 28% of the total text.

MOTOROLA TECHNICAL DEVELOPMENTS

INFORMATION SHEET

FOR DEFENSIVE PUBLICATION

Date: May 23, 2003

Title: Redundancy Implementation using OSPF weighting

Docket No.: CE11040W (17254)

Author #1

Name: Leonard K. Pennock

Location: iDEN 3G Development, GTSS

2501 South Price Rd

Chandler, AZ (AZ10)

 

Author #2

Name: Peter Emmons

Location: iDEN 3G Development, GTSS

2501 South Price Rd

Chandler, AZ (AZ10)


Redundancy Implementation using OSPF weighting

Leonard K. Pennock

Peter Emmons

I. Introduction

This paper outlines a method using OSPF to assign traffic to a favored server in a redundant server pool based on the states of the available servers and provides a method to avoid unwanted role reassignment.

II. Problem Solved

Traditional methods of routing traffic to a favored server in a server pool require communication between the network elements to determine which server is the best choice to be the active server. These methods generally have the concept of a primary server. This kind of configuration results in significant complexity. These methods generally have a problem with undesired traffic takeover when the “primary’ server is restarted or comes on line.

The method outlined in this paper allows for choosing the active server without having any direct server to server communication

. This results in increased isolation between servers, yielding significant availability improvements. The method also has prevention for undesired tariff takeover when a server is restarted.

III. How it Works

A.     Traffic stealing prevention

Open Shortest Path First (OSPF) routing protocol is based on the concept that traffic will take the lowest cost route when there are multiple paths available to get to the destination. This was originally designed to provide redundant routes to the same location with one of the routes being the favored route. By making two servers look like routes to the same logical IP, OSPF will allow two or more servers to have the same IP address and only the one with the lowest cost will receive traffic. This concept is not new and is used by most systems that use OSPF to manage redundancy.

Assuming there are two servers, the normal implementation will have a static configuration so that the primary server advertises a lower cost route then the secondary server advertises. Under normal operation the primary server will receive the traffic since it is the lower cost route. When the primary server fails the network detects the failure by the loss of OSPF Link State Announcements (LSA) from the server. When the LSA stops, the route cost to the primary server becomes infinite, and the secondary server becomes the lower cost route and starts receiving the traffic.

This handles failure of the server; however there is a problem when the failed server comes back on line. Since the costs are static, when the primary server comes back on line it starts advertising a lower cost than the secondary server. Since the primary server now has the lower cost, th...