Topologically-Aware Overlay Construction and Server SelectionSylvia Ratnasamy, Mark Handly, Richard Karp and Scott ShenkerPresented by Shreeram Sahasrabudhe (sas4@lehigh.edu) CSE 498 Advanced Networks

MotivationThe potential benefit by some knowledge of topology for distributed internet applicationsImproving application-level connectivityNeed for solution which is SimpleFast for the dynamic target systemsDistributed no central point of failure or bottleneckScalable for millions of nodes

ApproachNetwork Measurement used: Network LatencyNon-intrusiveLight-weightEnd-to-endPriorities: (Scalability + Practicality) > AccuracyBinning SchemeEvaluation of the application of this scheme to overlay networks and server selection

Distributed BinningRequires a set of landmark machines spread across the internet. (8-12 machines)Nodes measure RTT to each of these landmarks and orders the landmarks in increasing RTT. (relative)Divide the range of possible latency values into layers (absolute)

Scalable?Each node only needs measure with small set of landmarksAt a million nodes on the network, refreshing at every hour, each landmark would approximately handle 2700pings/sec. (800Mhz m/c; DNS root servers handle DNS queries at 1600/sec)Better scalability by have multiple nodes at a location act as a single logical landmark.

Sanity CheckFor each node in a bin computeGain ratio = inter-bin latency / intra-bin latency Ratio = Reduction in Latency = DesirableAlgorithm tested on simulated topologies (Transit Stub, Power-law Random Graphs) and internet data (NLANR). Improvement rapidly saturates Gain ratio is affected by the size of the underlying topology

Binning Vs (Random, Nearest-Neighbor)Random Binning: Each node selects a bin at random. Nearest Neighbor Clustering: At each iteration, two closest clusters are merged into a single cluster.

Construction of OverlaysStructured: Nodes are interconnected (at application-level) in a well-defined manner. Content-Addressable Network, Chord, PASTRY, TapestryUnstructured: Less structured networks End-system multicast, Scattercast.Measurement metric is Latency Stretch: ratio of average inter-node latency on the overlay network to the average inter-node latency on the underlying IP-level network. Latency Stretch = Better!

Construction of CAN topologies using BinningOrdering of landmarks is used for binningm landmarks, m! orderingsCo-ordinate space divided along first dimension into m portions, each portion sub divided along second dimension into m-1 portions and so onNew node joins CAN at a random portion associated with its landmark orderingResultCo-ordinate space not uniformly populatedUneven distribution of size of zone spaces (future work!)

Unstructured OverlaysGiven a set of n nodes on the Internet, have each node pick any k neighbor nodes from this set, so that the average routing latency on the resultant overlay is low.Use a heuristic algorithm: node picks k/2 closest nodes and k/2 at randomShort-Long: Shortest path routing, short-long-shortBinShort: k/2 nodes from self-bin and rest k/2 at random.

Server - SelectionIf (Server=bin[client]) > 1 Redirect to a random server in same binElse Select existing server with most_similar_bin to clients(Degree of similarity between two bins = no of matches of positions in landmark ordering)Stretch = (latency to selected server) / (latency to optimal server)

- For TS-10K 1000 servers, rest clients- Adjusted stretch- For NLANR data- Improved performance- diminishing returns