Please enable JavaScript.
Coggle requires JavaScript to display documents.
ECE454 (Fault-Tolerance (RAFT (Definition/Architecture, Leader Election,…
ECE454
Fault-Tolerance
Introduction
Dependability of a System
Type of Failure
Type of Faults
Masking by Redundancy
Consensus Problem
Apache Zookeeper
Distributed Commit and Checkpoints
2-phase Commit
State-transition
Recover From Failure
Deadlock of Participants
Algorithm
2-phase Checkpointing
Distributed Snapshot & Recovery Line
Algorithm
Problem/Motivation
RAFT
Definition/Architecture
Leader Election
Normal Operation
Recover From Failure & Fixing inconsistency
Architecture
Layer-based
Event-based
Object-based
Data-centered
P2P Systems
Chord Algorithm with Finger Table
CAP Principle
Problem with satisfying CAP
CAP definition
AP vs CP System
PACELC
Network Partition
Distributed File System
Network File System
File Sharing Semantics
Google File System
Client/Server interaction
Remote access model
Download/upload model
Introduction
Definition and motivation
Goals
Resource Sharing
Transparancy
Open & Scalable
Types of Distributed Systems
Communication
Remote Procedure Call (RPC)
RPC Steps
One-way, Sync, Async RPC
Apache Thrift
IDL
Message Queuing Model
Referential/Temporal Coupling
Transient vs Persistent Communication
Consistency and Replication
Motivation/Problem of Replicating Data
Consistency Models
Sequential consistency
Causal Consistency
Eventual Consistency with Session Guarantees
Monotonic Reads
Read your own writes
Linearizability
Replication Protocols
Primary Based Replication
Remote-write
Local-write
Quorum-based Protocols
Eventually-Consistent Protocol
Clocks
Lamport Clock
Update Algorithm
Happens-before relation
Vector Clocks Update Algorithm
Network Time Protocol
Stratum Definition
Delay Calculation
Offset Calculation
Big Data
Hadoop
Map/Reduce
Combiner
Spark
Graph Processing
Motivation
Google Pregel
BSP model with Superstep computation
PageRank
Single Source Shortest Path
Apache Kafka