02.10.2013 Views

FTOS Configuration Guide for the C-Series - Force10 Networks

FTOS Configuration Guide for the C-Series - Force10 Networks

FTOS Configuration Guide for the C-Series - Force10 Networks

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Chapter 30<br />

Bidirectional Forwarding Detection is supported only on plat<strong>for</strong>ms: c e<br />

Protocol Overview<br />

Bidirectional Forwarding<br />

Detection<br />

Bidirectional Forwarding Detection (BFD) is a protocol that is used to rapidly detect communication<br />

failures between two adjacent systems. It is a simple and lightweight replacement <strong>for</strong> existing routing<br />

protocol link state detection mechanisms. It also provides a failure detection solution <strong>for</strong> links on which no<br />

routing protocol is used.<br />

BFD is a simple hello mechanism. Two neighboring systems running BFD establish a session using a<br />

three-way handshake. After <strong>the</strong> session has been established, <strong>the</strong> systems exchange periodic control<br />

packets at sub-second intervals. If a system does not receive a hello packet within a specified amount of<br />

time, routing protocols are notified that <strong>the</strong> <strong>for</strong>warding path is down.<br />

BFD provides <strong>for</strong>warding path failure detection times on <strong>the</strong> order of milliseconds ra<strong>the</strong>r than seconds as<br />

with conventional routing protocol hellos. It is independent of routing protocols, and as such provides a<br />

consistent method of failure detection when used across a network. <strong>Networks</strong> converge faster because BFD<br />

triggers link state changes in <strong>the</strong> routing protocol sooner and more consistently, because BFD can<br />

eliminate <strong>the</strong> use of multiple protocol-dependent timers and methods.<br />

BFD also carries less overhead than routing protocol hello mechanisms. Control packets can be<br />

encapsulated in any <strong>for</strong>m that is convenient, and, on <strong>Force10</strong> routers, sessions are maintained by BFD<br />

Agents that reside on <strong>the</strong> line card, which frees resources on <strong>the</strong> RPM. Only session state changes are<br />

reported to <strong>the</strong> BFD Manager (on <strong>the</strong> RPM), which in turn notifies <strong>the</strong> routing protocols that are registered<br />

with it.<br />

BFD is an independent and generic protocol, which all media, topologies, and routing protocols can<br />

support using any encapsulation. <strong>Force10</strong> has implemented BFD at Layer 3 and with UDP encapsulation.<br />

BFD functionality will be implemented in phases. OSPF, IS-IS (not on C-<strong>Series</strong>), VRRP, VLANs, LAGs,<br />

static routes, and physical ports support BFD, based on <strong>the</strong> IETF internet draft draft-ietf-bfd-base-03.<br />

<strong>FTOS</strong> <strong>Configuration</strong> <strong>Guide</strong>, version 7.7.1.0 597

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!