<?xml version="1.0" encoding="UTF-8"?>
<reference anchor="I-D.choudhury-manral-flooding-simulation" target="https://datatracker.ietf.org/doc/html/draft-choudhury-manral-flooding-simulation-00">
   <front>
      <title>LSA Flooding Optimization Algorithms and Their Simulation Study
</title>
      <author initials="G." surname="Choudhury" fullname="G Choudhury">
         </author>
      <author initials="V." surname="Manral" fullname="Vishwas Manral">
         <organization>Netplane Systems</organization>
      </author>
      <date month="November" day="5" year="2002" />
      <abstract>
	 <t>The full flooding of LSAs in OSPF may cause large CPU and memory
consumption at node processors of a network with large number of nodes,
links, adjacencies per node and LSDB size.  An LSA storm, defined as  
the near-simultaneous update of a large number of LSAs, in such  
networks may cause network instability and outage.

We do a simulation study of four alternative algorithms to full 
flooding to determine their ability to handle large LSA storms.
Algorithm 2 does full flooding but in case two neighbors are connected
by multiple interfaces, flooding is done over only one such interface.
The other algorithms are based on Algorithm 2 and employ
further flooding restrictions. In Algorithm 3 each node asks only
a subset of its one-hop neighbors, known as multipoint relays, to
flood further. In Algorithm 4 flooding is done only over a
minimum spanning tree. Algorithm 5 uses full flooding (as in Algorithm
2) for LSAs carrying intra-area topology information and restricted
flooding over a minimum spanning tree (as in Algorithm 4) for other
LSAs.
	 </t>
      </abstract>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-choudhury-manral-flooding-simulation-00" />
   
</reference>
