23-04-2011, 03:44 PM
Presented By
Mangesh Chanrakant Swami
[attachment=12718]
Seminar on DUAL QUORUM REPLICATION SYSTEM FOR EDGE SERVICES
DEFINITIONS
What is Quorum?
What is Replication System?
What is Edge Service?
Introduction
Object stored on a single server.
But what if the server fails ?
Quorum System
Client needs a quorum of servers to access the object
What is a quorum system?
All quorums must have at least one server in common with all other quorums.
SYSTEM MODEL
front end nodes that handle service client requests from across the Internet, execute application-specific processing, and act as edge server clients or just clients to the dual-quorum storage system
Output Quorum System (OQS) nodes that process client read requests.
Input Quorum System (IQS) nodes that process client write requests.
OPERTION
WRITE:-
Client queries a quorum Q for a timestamp from each server in Q
Client chooses a timestamp t greater than any of those returned, and greater than its own timestamp
Client writes v by sending <v, t> to servers:
Each server updates v if t > v’s timestamp on the server
The server returns an ack regardless of whether it updated v
Once the client has received an ack from every server in some quorum Q’, the write is complete
Q does not have to be the same as Q’
Reads:-
Client queries every server in some quorum Q for value/timestamp pairs <v , t>
Client applies some deterministic function Result(A) to the returned pairs
Result(A) should return the correct value v to indicate a read failure
However, the details of Result(A) are intentionally left unspecified to allow different implementations of quorum systems
Dual-Quorum Protocol Design
The basic idea is to separate the read and write quorum into two quorum systems so that they can be optimized individually to improve response time and availability for read-dominated or write-dominated workloads.
The two quorum systems conditionally synchronize with each other to maintain the consistency of data replicated on them when processing both reads and writes.
The protocol is discussed in two steps:-
Simplified asynchronous DQ (ADQ) protocol:-
This protocol allows independent optimizations of read and writes quorums.
Volume leases to the protocol :-
This protocol used to improve writing availability while retaining good read performance.
Dual Quorum Protocol
Asynchronous Dual-Quorum Protocol:-
Basic read and writes operations
Read hit and read miss
Dual-Quorum with Volume Leases
Data structures
APPLICATION
This protocol used in WAN to access Internet with speed and consistency.
Quorum is also used for sensing in Genomics i.e. Quorum sensing.
Quorums can also be used in Grid Technology.
ADVANTAGES
It is a Fault Tolerance
It can handle concurrent operations.
Improves service availability and latency.
It offers nearly ideal trade-offs among high availability, good performance, and strong consistency.
DISADVANTAGE
Support for data replication is a key problem in realizing the promise of Internet edge services.
CONCLUSION
The Internet edge service architecture attempts to improve service availability and latency by allowing clients to access the closest available edge server rather than a centralized server or a centralized server cluster. To provide a single service from multiple locations, service logic (code) replicated on all edge servers must access a collection of shared data.
Mangesh Chanrakant Swami
[attachment=12718]
Seminar on DUAL QUORUM REPLICATION SYSTEM FOR EDGE SERVICES
DEFINITIONS
What is Quorum?
What is Replication System?
What is Edge Service?
Introduction
Object stored on a single server.
But what if the server fails ?
Quorum System
Client needs a quorum of servers to access the object
What is a quorum system?
All quorums must have at least one server in common with all other quorums.
SYSTEM MODEL
front end nodes that handle service client requests from across the Internet, execute application-specific processing, and act as edge server clients or just clients to the dual-quorum storage system
Output Quorum System (OQS) nodes that process client read requests.
Input Quorum System (IQS) nodes that process client write requests.
OPERTION
WRITE:-
Client queries a quorum Q for a timestamp from each server in Q
Client chooses a timestamp t greater than any of those returned, and greater than its own timestamp
Client writes v by sending <v, t> to servers:
Each server updates v if t > v’s timestamp on the server
The server returns an ack regardless of whether it updated v
Once the client has received an ack from every server in some quorum Q’, the write is complete
Q does not have to be the same as Q’
Reads:-
Client queries every server in some quorum Q for value/timestamp pairs <v , t>
Client applies some deterministic function Result(A) to the returned pairs
Result(A) should return the correct value v to indicate a read failure
However, the details of Result(A) are intentionally left unspecified to allow different implementations of quorum systems
Dual-Quorum Protocol Design
The basic idea is to separate the read and write quorum into two quorum systems so that they can be optimized individually to improve response time and availability for read-dominated or write-dominated workloads.
The two quorum systems conditionally synchronize with each other to maintain the consistency of data replicated on them when processing both reads and writes.
The protocol is discussed in two steps:-
Simplified asynchronous DQ (ADQ) protocol:-
This protocol allows independent optimizations of read and writes quorums.
Volume leases to the protocol :-
This protocol used to improve writing availability while retaining good read performance.
Dual Quorum Protocol
Asynchronous Dual-Quorum Protocol:-
Basic read and writes operations
Read hit and read miss
Dual-Quorum with Volume Leases
Data structures
APPLICATION
This protocol used in WAN to access Internet with speed and consistency.
Quorum is also used for sensing in Genomics i.e. Quorum sensing.
Quorums can also be used in Grid Technology.
ADVANTAGES
It is a Fault Tolerance
It can handle concurrent operations.
Improves service availability and latency.
It offers nearly ideal trade-offs among high availability, good performance, and strong consistency.
DISADVANTAGE
Support for data replication is a key problem in realizing the promise of Internet edge services.
CONCLUSION
The Internet edge service architecture attempts to improve service availability and latency by allowing clients to access the closest available edge server rather than a centralized server or a centralized server cluster. To provide a single service from multiple locations, service logic (code) replicated on all edge servers must access a collection of shared data.
