Browse Prior Art Database

Advanced Keep-Alive Procedure in Redundant Systems

IP.com Disclosure Number: IPCOM000124456D
Published in the IP.com Journal: Volume 5 Issue 5 (2005-05-20)
Included in the Prior Art Database: 2005-May-20
Document File: 4 page(s) / 402K

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

A transport system (packet network) should forward packets from the source system to the destination system in a redundant (primary and backup branch) way. It is a technical challenge to implement a transport system which on the one hand uses the system resources efficiently and on the other hand is redundant enough to handle possible transmission failures. Figure 1 illustrates the here referred structures and functions: Keep alive message handling in the primary branch: To ensure that outages shorter than the time between two keep alive messages of the primary sender are not causing a loss of alarms the primary sender has to send the keep-alive message after coming up only when the time period between one keep alive message is elapsed and not directly after it is coming up. See example 1, when the newly arrived packets are already forwarded. This is not needed when the source system is retrying to deliver the packets to the primary branch. With a small enhancement it is possible to handle some error conditions in the "underlying parts", primary receiver and LAN connection in the primary branch. If the sender is sending the keep-alive message, when at least one packet was received directly after a keep-alive message has been sent, refer to example 2. This handles the case when the error condition persists for at least double as long as the period between the keep-alive messages.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 45% of the total text.

Page 1 of 4

S

Advanced Keep-Alive Procedure in Redundant Systems

Idea: Sven Bindczeck, DE-Muenchen

A transport system (packet network) should forward packets from the source system to the destination system in a redundant (primary and backup branch) way. It is a technical challenge to implement a transport system which on the one hand uses the system resources efficiently and on the other hand is redundant enough to handle possible transmission failures.

Figure 1 illustrates the here referred structures and functions: Keep alive message handling in the primary branch: To ensure that outages shorter than the time between two keep alive messages of the primary sender are not causing a loss of alarms the primary sender has to send the keep-alive message after coming up only when the time period between one keep alive message is elapsed and not directly after it is coming up. See example 1, when the newly arrived packets are already forwarded. This is not needed when the source system is retrying to deliver the packets to the primary branch.

With a small enhancement it is possible to handle some error conditions in the "underlying parts", primary receiver and LAN connection in the primary branch. If the sender is sending the keep-alive message, when at least one packet was received directly after a keep-alive message has been sent, refer to example 2. This handles the case when the error condition persists for at least double as long as the period between the keep-alive messages.

Keep-alive message handling in the backup branch: When the backup branch receives a keep-alive message, it is removing all packets from its internal queue. Those were stored before the last keep- alive message was received. If the keep-alive message is not received, then it forwards all packets from the queue and also all newly received packets until the next keep-alive message from the primary branch is received (Example 3a, 3b).

Time interval between two keep-alive messages: The time interval between two keep-alive messages must be set bigger than the maximum delta processing time of the two branches plus the delay time in case of bursts (recovery time) and plus the maximum processing time for the retry mechanism of the primary branch (Example 4a, 4b). Figure 3 shows an example where the keep-alive time message interval was calculated too short (tA, tB and tC are arbitrary values).

Example 1:

- The primary branch is sending the keep-alive message - The primary branch is going down - The backup branch is receiving the packet X from its receiver and stores it internally - The primary branch is coming up - The primary branch is receiving the packet X+1 from its receiver and sends it - The backup branch is receiving the packet X+1 from its receiver and stores it internally - The time, based on the last keep-alive message has elapsed, but the primary did not send it, because it got down in between - The backup branch is not receiving the keep-alive message as expected and s...