- The document discusses various agreement protocols in distributed systems, including consensus algorithms for crash failures and Byzantine agreement problems.
- It presents the core problems in distributed systems like agreement, clock synchronization, stable storage, and reliable communication.
- Consensus algorithms for crash failures require f+1 rounds of message passing to tolerate f failures, while Byzantine agreement is impossible with less than 3f+1 processors or in asynchronous systems.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online from Scribd
- The document discusses various agreement protocols in distributed systems, including consensus algorithms for crash failures and Byzantine agreement problems.
- It presents the core problems in distributed systems like agreement, clock synchronization, stable storage, and reliable communication.
- Consensus algorithms for crash failures require f+1 rounds of message passing to tolerate f failures, while Byzantine agreement is impossible with less than 3f+1 processors or in asynchronous systems.
- The document discusses various agreement protocols in distributed systems, including consensus algorithms for crash failures and Byzantine agreement problems.
- It presents the core problems in distributed systems like agreement, clock synchronization, stable storage, and reliable communication.
- Consensus algorithms for crash failures require f+1 rounds of message passing to tolerate f failures, while Byzantine agreement is impossible with less than 3f+1 processors or in asynchronous systems.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online from Scribd
- The document discusses various agreement protocols in distributed systems, including consensus algorithms for crash failures and Byzantine agreement problems.
- It presents the core problems in distributed systems like agreement, clock synchronization, stable storage, and reliable communication.
- Consensus algorithms for crash failures require f+1 rounds of message passing to tolerate f failures, while Byzantine agreement is impossible with less than 3f+1 processors or in asynchronous systems.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online from Scribd
Download as doc, pdf, or txt
You are on page 1/ 21
Agreement Protocols
CS60002: Distributed Systems
Pallab Dasgupta Dept. of Computer Sc. & Engg., Indian Institute of Technology Kharagpur
Dept. of CSE, IIT KGP
Classification of Faults Based on components that failed Program / process Processor / machine Link Storage Clock
Based on behavior of faulty component
Crash just halts Failstop crash with additional conditions Omission fails to perform some steps Byzantine behaves arbitrarily Timing violates timing constraints
Dept. of CSE, IIT KGP
Classification of Tolerance Types of tolerance: Masking system always behaves as per specifications even in presence of faults Non-masking system may violate specifications in presence of faults. Should at least behave in a well-defined manner
Fault tolerant system should specify:
Class of faults tolerated What tolerance is given from each class
Dept. of CSE, IIT KGP
Core problems Agreement (multiple processes agree on some value) Clock synchronization Stable storage (data accessible after crash) Reliable communication (point-to-point, broadcast, multicast) Atomic
Dept. of CSE, IIT KGP
Overview of Consensus Results
Let f be the maximum number of faulty processors. Tight bounds for message passing: Crash failures Number of rounds Total number of processors Message size f+1 f+1 polynomial Byzantine failures f+1 3f + 1 polynomial
Dept. of CSE, IIT KGP
Overview of Consensus Results
Impossible in asynchronous case. Even if we only want to tolerate a single crash failure. True both for message passing and shared read-write memory.
Dept. of CSE, I KGP IIT
Consensus Algorithm for Crash Failures
Code for each processor: v := my input at each round 1 through f+1: if I have not yet sent v then send v to all wait to receive messages for this round v := minimum among all received values and current value of v if this is round f+1 then decide on v
Correctness of Crash Consensus Algo
Termination: By the code, finish in round f + 1.
Validity: Holds since processors do not introduce spurious messages
if all inputs are the same, then that is the only value ever in circulation.
Dept. of CSE, IIT KGP
Correctness of Crash Consensus Algo
Agreement: Suppose in contradiction pj decides on a smaller value, x, than does pi. Then x was hidden from pi by a chain of faulty processors: round round round f round f+1
q1
q2
qf
qf+1
pj pi
There are f + 1 faulty processors in this chain, a contradiction.
Dept. of CSE, IIT KGP
Performance of Crash Consensus Algo
Number of processors n > f f + 1 rounds n2 |V| messages, each of size log|V| bits, where V is the input set.
Dept. of CSE, I KGP IIT
Lower Bound on Rounds
Assumptions: n>f+1 every processor is supposed to send a message to every other processor in every round Input set is {0,1}
Dept. of CSE, I KGP IIT
Byzantine Agreement Problems
Model : Total of n processes, at most m of which can be faulty Reliable communication medium Fully connected Receiver always knows the identity of the sender of a message Byzantine faults Synchronous system In each round, a process receives messages, performs computation, and sends messages.
Dept. of CSE, IIT KGP
Byzantine Agreement
Also known as Byzantine Generals problem One process x broadcasts a value v
Agreement Condition: All non-faulty processes must agree on a common value. Validity Condition: The agreed upon value must be v if x is non-faulty.
Dept. of CSE, IIT KGP
Variants Consensus Each process broadcasts its initial value Satisfy agreement condition If initial value of all non-faulty processes is v, then the agreed upon value must be v
Interactive Consistency Each process k broadcasts its own value vk
All non-faulty processes agree on a common vector (v1,v2,,vn) If the kth process is non-faulty, then the kth value in the vector agreed upon by non-faultyprocesses must be vk
Solution to Byzantine agreement problem implies solution to other two
Dept. of CSE, IIT KGP
Byzantine Agreement Problem
No solution possible if: asynchronous system, or n < (3m + 1)
Lower Bound: Needs at least (m+1) rounds of message exchanges
Oral messages messages can be forged / changed in any manner, but the receiver always knows the sender
Dept. of CSE, IIT KGP
Proof Theorem: There is no t-Byzantine-robust broadcast protocol for t N/3
S 0 0 0 0 1 0 1 1
S S 1 0 1 1
U U
T T
U U
Scenario-0: T must decide 0
Scenario-1: U must decide 0
S S 0 0 1 0 1 1
T Dept. of CSE, IIT KGP
U U U
Scenario-2: -- similar to Scenario-0 for T -- similar to Scenario-1 for U -- T decides 0 and U decides 1
Lamport-Shostak-Pease Algorithm Algorithm Broadcast( N, t ) where t is the resilience For t = 0, Broadcast( N, 0 ): Pulse 1
The general sends value, xg to all processes, the lieutenants do not send. Receive messages of pulse 1. The general decides on xg. Lieutenants decide as follows: if a message value, x was received from g in pulse-1 then decide on x else decide on udef
Dept. of CSE, IIT KGP
Lamport-Shostak-Pease Algorithm contd..
For t > 0, Broadcast( N, t ): Pulse 1 The general sends value, xg to all processes, the lieutenants do not send. Receive messages of pulse 1. Lieutenant p acts as follows: if a message value, x was received from g in pulse-1 then xp = x else xp = udef ; Announce xp to the other lieutenants by acting as a general in Broadcastp( N 1, t 1 ) in the next pulse Pulse t +1 Receive messages of pulse t +1. The general decides on xg. For lieutenant p: A decision occurs in Broadcastq( N 1, t 1 ) for each lieutenant q Wp[q] = decision in Broadcastq( N 1, t 1 ) yp = max (Wp)
Dept. of CSE, IIT KGP
Features Termination: If Broadcast( N, t ) is started in pulse 1, every process decides in pulse t + 1 Dependence: If the general is correct, if there are f faulty processes, and if N > 2f + t, then all correct processes decide on the input of the general Agreement: All correct processes decide on the same value The Broadcast( N, t ) protocol is a t-Byzantine-robust broadcast protocol for t < N/3 Time complexity: O( t + 1 ) Dept. of CSE, IIT KGP
