Path Protection in MPLS Networks Using Segment Based Approach
Overview Short Intro to MPLS Introduction to Our work – Protection and the segment Based approach Algorithms for QoS constraints –Switch Over time Algorithm Greedy Approach Consideration of Backup Paths –Conserving protection resources – sharing bw –End-to-End delay and Jitter –Combining the above constraints –Reliability Visualization System Experimental Results Conclusion and work done
Introduction to MPLS Mapping: 0.40 Request: 47.1 Mapping: 0.50 Request: 47.1
Label Switched Path (LSP) IP
MPLS : ROUTE AT EDGE, SWITCH IN CORE IP Forwarding LABEL SWITCHING IP Forwarding IP #L1IP#L2IP#L3 IP Applies concept of VC routing Packet forwarding is done based on Label Switching FEC: Destination address prefix, Traffic Engineering tunnel, Class of Service.
Introduction to Path Protection
Introduction to Path Protection BACKUP PATH
Requirements of Path Protection Should Reroute the traffic satisfying certain QoS constraints Should aim to conserve the amount of protection resources reserved
Global Path Protection Backup Path
Local Path Protection
Segment Based Path Protection Look at the path as a group of segments – protect each segment separately Results in fewer backup paths – conserves resources Meets QoS constraints in a “tight” manner Gives flexibility Issue : How to segment the path ?
Algorithms for QoS constraints
QoS Constraints Important parameters –Switch-Over Time –End-to-End Delay –Jitter –Reliability –Combination of above Have to conserve protection resources
Bounded Switch Over Time Definition of Switch Over Time
An expression for switch over time Analysis for switch over time RTT( R i, R j ) + T test <
Example for Segment Based Approach
Here we are able to meet the Switch Over time constraint with 3 backup paths as compared to 7 backup paths in LPP A simple algorithm for segmentation: Greedy Approach
The Resource advantage
Problem with Greedy Approach Need to consider the topology of the network as well
An adaptive Algorithm for segmentation Start from the egress and look for longest possible segment
End-to-End Delay An important parameter
Analysis Max (T + ( t2 – t1 ) ) < EED Bound
Algorithm for end-to-end delay For each backup path, we need to make sure that the end-to-end constraint is satisfied Use shortest path approach for finding a backup path – minimizes end-to-end delay
Algorithm for end-to-end delay d1d1 d2d2 d3d3 d 1 + d 2 + d 3 d3d3 0 d 2 + d 3 Searching for a backup path
Jitter Jitter can be treated as a link property Path Jitter = Σ Link Jitter Algorithm similar to end-to-end delay
Combination of above constraints Approach Dynamic Programming Switch Over Time End-to-End Delay Jitter A combined Algorithm for
Algorithm based on Dynamic Programming Artificial Node RiRi RjRj RkRk
Reliability An important QoS parameter in Computer Networks. Path Reliability : Probability of a path to be in a working state at some instant of time. Link Reliability (p) : Probability of a link to be in working state at some instant of time.
Reliability - Objectives Effect of Path Protection on Reliability Effect of Segment Size on Reliability An O(No. of Links + (No. of Segments) 2 ) Algorithm to find exact path reliability ! Algorithm for Finding most reliable Backup Path Heuristics for SBPP with reliability bounds
Effect of Path Protection on Reliability Total number of links in primary path = n Reliability of a link : p Path Reliability from A to B = p n A B Path Reliability from A to B with backup path = 2p n – p 2n n links
Effect of Path Protection on Reliability
Effect of Segment Size on Reliability Total number of links in primary path = n Reliability of a link : p Size of Segments = k Number of Segments = n/k Size of Backup Path = Size of Segment Reliability of the path = (2p k – p 2k ) n/k
Effect of Segment Size on Reliability
Algorithm to find path reliability Theoretically a path exists between ingress and egress nodes : R1 -> R2 -> R4 -> R5 -> R6 -> R7 No path between ingress and egress nodes in our path switching approach !
Algorithm to find path reliability Probability of Primary path for a particular segment S i to be working Probability of Backup path for Segment S i to be working Probability of S i to S j-1 segments’s primary path to be working & segment S j primary path to have an error
Algorithm for Finding most reliable Backup Path Artificial Node RiRi RjRj
Heuristics for SBPP with reliability bounds Divide any segment into two till the reliability bound is met Find the Segmentation with least number of segments
Visualization System
Visualization of Algorithms A visualization system developed based on POLKA – an algorithm animation toolkit Closely Integrated with the simulator Aids in understanding how the algorithms work Assists in establishing correctness of algorithms and simulations Dynamic Nature of Visualizations
Visualization Two categories of Visualization: –Animation for Adaptive Bounded Switch Over Time Algorithm –Rerouting of packets by SSR in case of failure (packet flow animation) – demonstrates various cases
Topology for Visualization
Visualization Demo
Experimental Results
Implementation Simulator developed in C++ for implemented some algorithms Size of model graph : 100 nodes, 1000 edges RTT of each link = 10 ms BW – 50 to 100 Generated large number of random LSP requests and observed various parameters Results indicate advantages of SBPP
Segment Size vs BW reserved
Segment Size vs Rejection Rate ( for 250 LSPs )
No. of Requested LSPs vs Rejection Rate
Effect of Backup Path Sharing
Crossover - Effects of backup path sharing
Effect of additional constraint
Bandwidth Reserved vs Density
Detection and Notification
A Mechanism for Notification After a fault is detected, notification needs to be sent to the SSR for switching the traffic Some nodes will participate in notification and the SSR will switch the route What information will be passed after a fault occurs ? What changes do we need in the LSR tables for switching? Case of Multiple LSPs : All LSPs using that segment may not pass through the faulty node/link – Only concerned LSPs should be switched
A Mechanism for Notification
Conclusions Fault Tolerance can now be assured to satisfy various QoS constraints Segment Based Algorithms show significant improvement in terms of protection resources used
Work Done Mechanisms for Detection, Notification Algorithms for various QoS constraints –Bounded Switch over time –End-to-End delay –Jitter –Combination of above –Reliability Issues relating to backup path - sharing Simulator developed for above Visualization System