Responsive Distributed Routing Algorithm
- 1. Literature Review on “A Responsive Distributed Routing Algorithm for Computer Networks” JEFFREY M. JAFFE AND FRANKLIN H. MOSS • Nafiz Ishtiaque Ahmed •Rafi Al Razi •Mohammed Rokon Uddin •
- 2. Introduction • Each node uses information received from neighbors to determine its best next node through “summery” . • In distributer network “summery”(routing table) changes periodically . • This change causes temporary loops for a package . • On going package failed if the hop of existing package don’t has link to reach DES in its new routing table .
- 3. Motivation Analysis a new dynamic distributed algorithm for • Temporary Loops free routing . • Take Less Recovery Time(h) than previous algorithm (h^2) when Network Links fails .
- 4. Research Proposal • Maintains “summary” information rather than global topology by IUP . • A hop count field is included in the messages and tables for IUP to reasonably fast detection of loops . • Uptree nodes can update link cost but can’t change on going package route that means Uptree nodes are “Frozen” till it receives DES ack for “Unfrozen” .
- 5. Research Proposal • Multiple freeze ack are processed in FIFO order . • If a uptree node frozen then new link cost have to wait to entry to the table till it receives unfrozen ACK . • That means it takes O(x)time in worst case for unfrozen all nodes and O(h)time for sub- tree nodes to recovery .
- 6. Finding • Take more steps to recovery and still O(x) and O(h) is a hues time for a hues network . • Each Hop has to remember hues “summery table” and run through algorithm so it need memory and it takes more time to delivery . • Multiple frozen route in a hop can case noisy jam in a router thus no package can pass through the hop .
- 7. Solution • Global routing topology may use to solve this problem . • Package may used to memorize path in Global routing rather being “frozen” a hop .
- 8. Conclusion • This algorithm was developed for loop- free paths in the circumstance of changing link weights and minimize recovery time . • Modern Dynamic distributed algorithm was rooted on this algorithm .
Editor's Notes
- #3: 1. each node periodically sends “summary” information to its neighbors regarding its current best cost to one or more destinations. Each node uses information received from neighbors to determine its best next nodes.One basic attribute of these algorithms is that there is no notion of exercising control over the ordering in which the nodes perform their updates.