Presentation is loading. Please wait.

Presentation is loading. Please wait.

Nicolas Schiper Fernando Pedone University of Lugano Switzerland

Similar presentations


Presentation on theme: "Nicolas Schiper Fernando Pedone University of Lugano Switzerland"— Presentation transcript:

1 Nicolas Schiper Fernando Pedone University of Lugano Switzerland
Optimal Atomic Broadcast and Multicast Algorithms for Wide Area Networks Nicolas Schiper Fernando Pedone University of Lugano Switzerland

2 Atomic Broadcast and Multicast
Atomic Multicast: Atomic Broadcast: m2 m3 m1 m1 server 4 server 1 server 2 server 3 server 5 server 6 g1 AMCast m2 to g2, g3 ABCast m1 AMCast m1 to g1, g2 m2 m1 ABCast m2 g2 m2 m1 m2 AMCast m3 to all groups ABCast m3 m2 m1 g3 m2 m1 m1 < m2 <=> ∃p : p A-Delivers m1 before m2 The relation < is acyclic m2 m1

3 Problem (2) Atomic multicast is reducible to atomic broadcast:
Every multicast msg. is broadcast, only addresses of the msg. deliver it Solution of no practical interest: Communication among processes unrelated to msg. Genuine atomic multicast: Only processes addressed by the message take part in the protocol

4 Context Wide Area Networks: Toronto Melbourne Lugano Beijing
Fast and cheap Toronto Lugano Beijing Melbourne Slow and expensive Repl locally for locality and globally to tolerate catastrophic failures

5 Goal Minimize number of inter-group msg. delays between multicast and deliver A-Del(m) A-XCast(m) Database sites should only handle transactions that touch data items they replicate => 2 inter-group msg. delays for delivery

6 In The Full Paper (1) … Genuine atomic multicast lower bound:
Minimum 2 inter-group msg. delay We present an algorithm that achieves this lower bound Based on Skeen’s multicast algorithm (not fault-tolerant) Optimized version of Fritzke et al’s. (SRDS’98)

7 In The Full Paper (2) … Atomic Broadcast:
We present an atomic broadcast algorithm that needs only 1 inter-group msg. delay The algorithm is proactive, nevertheless quiescent Msg. broadcast after algo. quiescent => inter-group msg. delay for delivery Quiescence has an inherent cost: Our algorithm is optimal

8 Final Remarks Latency degree vs. msg. complexity tradeoff for data replication: Update data item A Optimize latency => atomic broadcast … but high msg. complex. Optimize msg. complex. => atomic multicast … but at least 2 inter-group msg. delays A B, C A,B A, B D, B


Download ppt "Nicolas Schiper Fernando Pedone University of Lugano Switzerland"

Similar presentations


Ads by Google