1. Pastry Antony Rowstron and Peter Druschel Presented By David Deschenes

2. Introduction General substrate for P2P applications General substrate for P2P applications Provides object location and routing Provides object location and routing Characteristics Characteristics Completely decentralized Completely decentralized Fault-resilient Fault-resilient Scalable Scalable Reliability Reliability

3. Design Circular Node Space Circular Node Space Nodes randomly assigned unique 128 bit identifiers Nodes randomly assigned unique 128 bit identifiers Uniform distribution of identifiers Uniform distribution of identifiers Identifiers indicate position in space Identifiers indicate position in space Messages and Keys Messages and Keys Messages assigned N-bit keys Messages assigned N-bit keys Each message routed to the node whose identifier is numerically closest to its assigned key Each message routed to the node whose identifier is numerically closest to its assigned key Key generation is application specific Key generation is application specific

4. Node State Leaf Set Leaf Set L closest nodes in the node space L closest nodes in the node space Routing Table Routing Table Identifies 2 b -1 hosts whose identifiers match the current node identifier in n bits Identifies 2 b -1 hosts whose identifiers match the current node identifier in n bits Neighborhood Set Neighborhood Set M closest nodes according to a proximity metric M closest nodes according to a proximity metric

5. Network Organization Node Arrival Node Arrival New node acquires the identifier X and routes a join message with key X through node A New node acquires the identifier X and routes a join message with key X through node A New nodes initialize their state based upon state shared by nodes that route its join message New nodes initialize their state based upon state shared by nodes that route its join message Node Departure Node Departure If a member of a node’s leaf set is lost, the leaf set is expanded outward to find a new node If a member of a node’s leaf set is lost, the leaf set is expanded outward to find a new node If a routing table entry is lost a node in the same row is consulted for a replacement If a routing table entry is lost a node in the same row is consulted for a replacement

6. Maintaining Locality Proximity based on a scalar metric Proximity based on a scalar metric E.g. IP routing hops, geographic distance E.g. IP routing hops, geographic distance Metric determined by application Metric determined by application Assumed that the triangulation inequality holds for the metric Assumed that the triangulation inequality holds for the metric Connections to nodes with greater proximity are desired Connections to nodes with greater proximity are desired Facilitated by the neighborhood set Facilitated by the neighborhood set

7. Message Routing Check the leaf set Check the leaf set If destination present route to it directly If destination present route to it directly Consult the routing table Consult the routing table Forward to a node whose identifier matches the message key in at least one more bit Forward to a node whose identifier matches the message key in at least one more bit Forward to a node whose identifier is numerically closer to the message key Forward to a node whose identifier is numerically closer to the message key

8. Evaluation Number of routing hops scales with the network size (log 2 b N) – Figure 4. Number of routing hops scales with the network size (log 2 b N) – Figure 4. Routes are only 30% to 40% longer despite minimal routing tables – Figure 6. Routes are only 30% to 40% longer despite minimal routing tables – Figure 6. Able to locate one of the top two nodes 92% of the time Able to locate one of the top two nodes 92% of the time Minimally impacted by node failure – Figure 10. Minimally impacted by node failure – Figure 10.