This was my B.E 7th Semester seminar report on byzantine fault

1. Synchronous Mortal Byzantine
A Fault Tolerant Mechanism For Distributed
Systems
Seminar Report By:
Bibaswann Bandyopadhyay
CSE, 7th Semester
Roll No. 20081017
University Institute of Technology

2. Introduction
 Several divisions of the Byzantine army are camped outside an
enemy city, each division commanded by its own general.
However, some of the generals may be traitors.
 Since the Byzantine Empire ceased to exists, we observe that
Byzantines were mortal, their traitors doubly so. In what
follows, we treat them as such.

3. Faults
 Deviation from expected behavior
 Variety of factors
 hardware
 software
 operator
 Network
 Three categories
 transient faults
 intermittent faults
 permanent faults
 Any fault may be
 fail-silent (fail-stop)
 Byzantine

4. What is a Byzantine Failure?
 Three primary differences from Fail-Stop Failure
1) Component can produce arbitrary output
• Fail-stop: produces correct output or none
2) Cannot always detect output is faulty
• Fail-stop: can always detect that component has stopped
3) Components may work together maliciously
• No collusion across components

5. Motivation
 Build reliable systems in the presence of faulty components
 Common approach:
 Have multiple (potentially faulty) components compute same function
 Perform majority vote on outputs to get “right” result
C1
C2 majority(v1,v2,v3)
C3
f faulty, f+1 good components ==> 2f+1 total

6. Key Step
A commanding general must send an order to his n-1 lieutenant
generals such that
IC1. All loyal lieutenants obey the same order.
IC2. If the commanding general is loyal, then every loyal
lieutenant obeys the order he sends.

7. Option 1: Loyal Commander
commander
attack
attack
L1 L2
retreat
What must L1 do?
By IC2: L1 must obey commander and attack

8. Option 2: Loyal L2
commander
retreat
attack
L1 L2
retreat
What must L1 do?
By IC1: L1 and L2 must obey same order --> L1 must retreat
Problem: L1 can’t distinguish between 2 scenarios

9. Oral Message Algorithm
 OM(0)
 Commander sends his value to every lieutenant
 OM(m), m>0
 Commander sends his value to every lieutenant
 For each i, let vi be value Lieutenant i receives from
commander; act as commander for OM(m-1) and send vi to
n-2 other lieutenants
 For each i and each j i, let vj be value Lieutenant i
received from Lieutenant j. Lieutenant i computes
majority(v1,...,vn-1)

10. Example: Bad Lieutenant
 Scenario: m=1, n=4, traitor = L3
A C A
OM(1):
A
L1 L2 L3
C
OM(0):???
A L2 L3
L1
R
A
R
Decision?? L1 = m (A, A, R); L2 = m (A, A, R); Both attack!

11. Example: Bad Commander
 Scenario: m=1, n=4, traitor = C
A C A
OM(1):
R
L1 L2 L3
A
OM(0):???
A R
L1 L2 L3
A
R
A
Decision?? L1=m(A, R, A); L2=m(A, R, A); L3=m(A,R,A); Attack!

12. Three Phase Protocol
 state of each replica is stored in a message log.
 Primary p receives a client request m , it starts a three-phase
protocol.
 Three phases are: pre-prepare, prepare, commit.
 Pre-prepare and prepare phases is used to totally order
requests.
 In pre-prepare phase
 Primary assigns sequence number n to request.
 Multicast pre-prepare msg. to all backups and appends the msg. to its
log.

13. Three phase protocol(contd.)

14. Conclusion
 The algorithm works correctly in asynchronous system like the internet.
 Previous algorithms are too slow to be used in practical (proportional to
the number of faulty nodes vs. number of phases)
 One reason why Byzantine fault tolerant algorithms is important in future
is that they allow the system to work correctly even when there are
software errors.
 not all, software errors that occur in all replicas
 It can mask errors that occur independently at different replicas
 Non-deterministic software errors
 Persistent errors

15. Thank you