09-P2P系统.ppt

3. Structured P2P: DHT Approaches 118 □ DHT service and issues □ CARP □ Consistent Hashing □ Chord □ CAN □ Pastry Pastry (in truth) 123 □ Nodes are assigned a 128-bit identifier □ The identifier is viewed in base 16 ○ e.g., 65a1fc04 ○ 16 subgroups for each group □ Each node maintains a routing table and a leaf set ○ routing table provides delegate nodes in nested groups ○ inner group idea flawed: might be empty or have too many nodes Pastry: Routing procedure if (destination is within range of our leaf set) forward to numerically closest member else if (there’s a longer prefix match in table) forward to node with longest match else forward to node in table (a) shares at least as long a prefix (b) is numerically closer than this node 125 Pastry: Performance Integrity of overlay/ message delivery: guaranteed unless L/2 simultaneous failures of nodes with adjacent nodeIds Number of routing hops: No failures: log16 N expected During failure recovery: O(N) worst case, average case much better 126 127 Pastry: Experimental results Prototype implemented in Java deployed testbed (currently ~25 sites worldwide) Simulations for large networks Pastry: Proximity routing Assumption: scalar proximity metric e.g. ping delay, # IP hops a node can probe distance to any other node Proximity invariant: Each routing table entry refers to a node close to the local node (in the proximity space), among all nodes with the appropriate nodeId prefix. 131 Pastry: Routes in proximity space 132 134 Pastry: Locality properties 1) Expected distance traveled by a message in the proximity space is within a small constant of the minimum 2) Routes of messages sent by nearby nodes with same keys converge at a node near the source nodes 3) Among k nodes with nodeIds closest to the key, message likely to reach the node closest to the source node first 136 Pastry: Summary O(log N) routing steps (expected) O(log N) routing table size Network proximity routing END * 哈希结果的变化意味着当缓冲空间发生变化时，所有