New version page

Fault Recovery

Upgrade to remove ads

This preview shows page 1-2-3-4 out of 11 pages.

Save
View Full Document
Premium Document
Do you want full access? Go Premium and unlock all 11 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 11 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 11 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 11 pages.
Access to all documents
Download any document
Ad free experience

Upgrade to remove ads
Unformatted text preview:

1IP Multicast Fault Recovery in PIM over OSPFAbstract—Relatively little attention has been given to understanding the fault recov-ery characteristics and performance tuning of native IP multicast networks.This paper focuses on the interaction of the component protocols to under-stand their behavior in network failure and recovery scenarios. We considera multicast environment based on the Protocol Independent Multicast (PIM)routing protocol, the Internet Group Management Protocol (IGMP) and theOpen Shortest Path First (OSPF) protocol. Analytical models are presentedto describe the interplay of all of these protocols in various multicast channelrecovery scenarios. Quantitative results for the recovery time of IP multicastchannels are given as references for network configurations, and protocol de-velopment. Simulation models are developed using the OPNET simulation toolto measure the fault recovery time and the associated protocol control over-head, and study the influence of important protocol parameters. A testbedwith five Cisco routers is configured with PIM, OSPF, and IGMP to measurethe multicast channel failure and recovery times for a variety of different linkand router failures. In general, the failure recovery is found to be light-weightin terms of control overhead and recovery time. Failure recovery time in aWAN is found to be dominated by the unicast protocol recovery process. Fail-ure recovery in a LAN is more complex, and strongly influenced by protocolinteractions and implementation specifics. Suggestions for improvement of thefailure recovery time via protocol enhancements, parameter tuning, and net-work configuration are provided.I. INTRODUCTIONMany IP multicast applications, for example, near real-time dis-semination of financial information, require high availability. Thisproblem has not received much attention so far. One exception isSTRESS [1], a tool that automates the formal evaluation of PIMsparse-mode protocol robustness. However, STRESS does not in-clude timers, and does not consider the interaction between unicastand multicast protocols.Multicast group membership management, unicast routing pro-tocols, and multicast routing protocols are all required to enableend-to-end multicast capabilities. In this paper, we investigate acomplete multicast routing architecture consisting of IGMP [6] formulticast group membership management in a LAN, OSPF [4] forunicast routing, and PIM sparse-mode [8] and PIM dense-mode[7] for multicast routing. OSPF is chosen because of its rapidfault recovery properties, widespread use, and its support of para-metrically tuning of fault recovery time, as compared with RIPwhich has long, non-tunable fail-over periods. The two variants ofPIM are becoming the dominant multicast routing protocols. Othermulticast protocols, such as DVMRP or CBT resemble dense andsparse mode, respectively, and we thus expect that many of ourresults apply to these and similar protocols as well. End-to-endmulticast channel fault recover is a function of the interplay of allof these protocols and is thus the focus of this paper.We investigate how quickly the multicast channel recovers whenlinks and routers fail in a multicast network. We define a multicastchannel as the state established in routers and hosts that allowsa single sender to communicate with a group of receivers. Weconsider single link and router faults inside the network, but weassume that sending and receiving hosts, their LANs are reliable.Since fault recoveryassociated with rendezvous point (RP) failuresin PIM SM have been studied extensively [8], this paperfocuses onother mechanisms (router, link, LAN, WAN fail-over) that are notsufficiently addressed and are less well understood by the commu-nity.The key aims of this study are: develop a detailed understandingof the protocol interactions and sequence of events under differentfailure scenarios; provide quantitative insight into the effect of pro-tocol parameters on recovery time and overhead; develop generalsuggestions for parametric tuning of protocols and enhancementsto protocol specifications and implementation. To achieve theseobjectives, we combine results from analytical analysis, simula-tions and testbed measurements.In the analysis, we present the interactions of the protocols(PIM, OSPF, IGMP) with end-to-end multicast channel recoveryunder various network failure scenarios. We also develop somequantitative results that can be used as references for network con-figurations and protocol development. In addition, the analysisserves as a basis for our providing recommendations on the pro-tocol enhancement.Simulation models for IGMP, PIM DM and support tools wereconstructed using the OPNET [11] simulation platform. The sim-ulation is used to measure the control costs of the trio of protocolsduring steady state and failure recovery scenarios, for various ran-dom topologies and with various parametric tunings. Furthermore,the simulation is used to validate the failure recovery results de-rived from the analytical models.The experimentalresults were supplemented by studying the op-eration and failure recovery of the protocols on a testbed of fiveCisco routers arranged in a simple topology. This enabled a basicdemonstration of failure recovery on WAN and LAN, and also al-lowed us to identify some implementation-relatedissues that affectfailure recovery.The paper is organized as follows. Section II reviews IGMP,OSPF and PIM. Section III describes the topologies and configu-rations we used, as well as the chain of events caused by link orrouter failures. We also present several analytical multicast recov-ery models. Section IV and V present the simulation and testbedresults, respectively.II. OVERVIEW OF PROTOCOLSThe establishmentof end-to-endmulticast communicationchan-nels requires several protocols to work in concert. To establish amulticast channel over a native multicast enabled WAN, a senderapplication needs only to send UDP packets onto the LAN usinga class D IP address (group address) in the destination field of theIP header. Multicast group information on a LAN is usually main-tained by the IGMP protocol. The multicast enabled routers in thenetworkare responsiblefor constructingthe multicast channel, andextending it to the interested receivers; in our case, this is done us-ing PIM DM or PIM SM. The multicast protocol constructs themulticast delivery tree using the metrics and topology informationmaintained by the unicast routing


Download Fault Recovery
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Fault Recovery and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Fault Recovery 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?