is cluster a peer to peer to peer system? Or do we use peer to peer system in a cluster?
---------------------------------
gossip-based membership
------------------------------------
Lamport timestamp
It follows some simple rules:
- A process increments its counter before each event in that process;
- When a process sends a message, it includes its counter value with the message;
- On receiving a message, the receiver process sets its counter to be greater than the maximum of its own value and the received value before it considers the message received.
Cassandra
http://www.datastax.com/documentation/articles/cassandra/cassandrathenandnow.html
virtual node in Cassandra
http://www.datastax.com/dev/blog/virtual-nodes-in-cassandra-1-2
column ~= cell, column family ~= table
consistent hashing
2 challenges:
First, the random position assignment of each node on the ring leads to non-uniform data and load distribution. Second, the basic algorithm is oblivious to the heterogeneity in the performance of nodes
solutions:
One is for nodes to get assigned to multiple positions in the circle (like in Dynamo), and the second is to analyze load information on the ring and have lightly loaded nodes move on the ring to alleviate heavily loaded nodes as described in
-----------------------------
CAP theorem
http://en.wikipedia.org/wiki/CAP_theorem
------------------------------------------------------------
Chandy–Lamport algorithm
-----------------------------------------------------
Virtual synchrony is an interprocess message passing (sometimes called ordered, reliable multicast) technology. Virtual synchrony systems allow programs running in a network to organize themselves into process groups, and to send messages to groups (as opposed to sending them to specific processes). Each message is delivered to all the group members, in the identical order, and this is true even when two messages are transmitted simultaneously by different senders. Application design and implementation is greatly simplified by this property: every group member sees the same events (group membership changes and incoming messages) in the same order.
http://en.wikipedia.org/wiki/Virtual_synchrony
--------------------------------------------------------------------
Bloom filter
http://stackoverflow.com/questions/4282375/what-is-the-advantage-to-using-bloom-filters
So an example usage pattern might be:
You've got a lot of data, on disk -- you decide on what error bound you want (e.g. 1%), that prescribes the value of m. Then the optimal k is determined (from the formula given in the article). You populate your filter from this disk-bound data once.
Now you have the filter in RAM. When you need to process some element, you query your filter to see if it stands a chance of existing in your data set. If it doesn't, no extra work is done. No disk reads, etc. (Which you would have to do if it were a hash or tree, etc).
Otherwise, if the filter says "Yes, it's in there", there's a 1% chance that it's wrong, so you do the necessary work to find out. 99% of the time, it really will be there, so the work was not for naught.
week1:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know the key differences between cloud computing and previous generations of distributed systems.
- Design MapReduce programs for a variety of problems.
- Know how Hadoop schedules jobs.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Clouds
- MapReduce paradigm
- Hadoop YARN
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- Why is cloud computing popular today?
- What is different in cloud computing compared to previous generations of distributed systems?
- How does one program in MapReduce?
- How does the MapReduce system schedule jobs?
Readings and Resources
No required reading for this week. But you could look at the Apache Hadoop documentation.
week2:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Analyze various gossip/epidemic protocols.
- Design and analyze various distributed membership protocols.
- Know what grid computing is.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you interact with the lectures. These topics will help you better understand the content in this module.- Failure detectors
- Membership protocols
- Gossip/epidemic protocols
- Grid computing
Guiding Questions
Develop your answers to the following guiding questions while completing the activities throughout the week.- Why are gossip and epidemic protocols fast and reliable?
- What is the most efficient way for cloud computing systems to detect failures of servers?
- How is grid computing related to cloud computing?
week3:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know how Napster, Gnutella, FastTrack, and BitTorrent work.
- Know and analyze how distributed hash tables work (Chord, Pastry, and Kelips).
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you interact with the lectures. These topics will help you better understand the content in this module.- Peer-to-peer systems
- Industrial P2P systems: Napster, Gnutella, FastTrack, BitTorrent
- Distributed hash tables: Chord, Pastry, Kelips
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- What is the difference between how Napster clients and Gnutella clients search for files?
- What is the difference between Gnutella and FastTrack?
- What is BitTorrent’s tit for tat mechanism?
- What is consistent hashing?
- Why are DHTs efficient in searching?
- How does Chord route queries?
- How does Pastry route queries?
- How does Kelips route queries?
- What is churn in P2P systems?
- How does Chord maintain correct neighbors in spite of failures and churn?
Readings and Resources
No readings are required for this week. But you can look at the following documentation:week 4:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know why key-value/NoSQL are gaining popularity
- Know the design of Apache Cassandra
- Know the design of Apache HBase
- Use various time synchronization algorithms
- Apply Lamport and vector timestamps to order events in a distributed system
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Key-value and NoSQL stores
- Cassandra system
- CAP theorem
- Consistency-availability tradeoff and spectrum
- Eventual consistency
- HBase system
- ACID vs. BASE
- Time synchronization algorithms in asynchronous systems: Cristian's, NTP, and Berkeley algorithms
- Lamport causality and timestamps
- Vector timestamps
Guiding Questions
Develop your answers to the following guiding questions while completing the assignments throughout the week.- Why are key-value/NoSQL systems popular today?
- How does Cassandra make writes fast?
- How does Cassandra handle failures?
- What is the CAP theorem?
- What is eventual consistency?
- What is a quorum?
- What are the different consistency levels in Cassandra?
- How do snitches work in Cassandra?
- Why is time synchronization hard in asynchronous systems?
- How can you reduce the error while synchronizing time across two machines over a network?
- How does HBase ensure consistency?
- What is Lamport causality?
- Can you assign Lamport timestamps to a run?
- Can you assign vector timestamps to a run?
Readings and Resources
There are no readings required for this week. But you can look at the following documentation:week 5
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Design an algorithm to calculate a distributed snapshot.
- Assign FIFO/Causal/Total ordering to multicast messages.
- Design a reliable multicast protocol.
- Know the working of the industry-standard protocol called Paxos.
- Know why consensus is hard to solve.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Global Snapshots
- Multicast Ordering
- Multicast Reliability
- Paxos
- Impossibility of Consensus
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- What is the difference between a safety property and a liveness property?
- How does the Chandy-Lamport algorithm work?
- How do you assign FIFO/Causal timestamps to multicasts in a distributed system?
- How does Paxos use quorums to ensure safety?
- Why is consensus impossible to solve in asynchronous systems?
Readings & Resources
No readings required for this week. But you can look at the following documentation:Part 2:
two-phase locking occurs inside of a transaction
release the lock when transaction reaches the commit point
week1:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Design Leader Election algorithms including Ring algorithm and Bully algorithm.
- Design Mutual Exclusion algorithms including Ricart-Agrawala’s algorithm and Maekawa’s algorithm.
- Know the design of Leader Election used in industry systems: Google’s Chubby system and Apache Zookeeper.
- Know how industry systems like Google’s Chubby support mutual exclusion.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Google Chubby Leader Election
- Apache Zookeeper Leader Election
- Ring Mutual Exclusion
- Ricart-Agrawala’s Mutual Exclusion
- Maekawa Mutual Exclusion
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- What are the safety and liveness conditions for Leader Election?
- Why is Leader Election hard?
- How long does the Ring Election algorithm take to complete?
- How long does the Bully Election algorithm take to complete?
- How does Google Chubby use quorums for election?
- What are the safety and liveness conditions for Mutual Exclusion?
- How do semaphores work?
- How long does the Ring Mutual Exclusion algorithm take to complete?
- How long does the Ricart-Agrawala’s algorithm take to complete?
- Why is Maekawa’s algorithm “optimal”?
- How does Google Chubby use quorums for mutual exclusion?
Readings & Resources
- (Optional, fun video) Amazon Web Services (AWS) Training Videos
- (Optional, fun video) Amazon Web Services (AWS) Training Videos (China host)
week2:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know how Remote Procedure Calls (RPCs) work.
- Check a run of transactions for correctness (serial equivalence).
- Design systems that use optimistic or pessimistic approaches to ensure correctness in spite of many concurrent clients.
- Detect and avoid deadlocks.
- Calculate nines availability for a replicated system.
- Know how to ensure correctness (consistency) in spite of multiple servers.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- LPCs vs RPCs
- Marshaling
- Serial Equivalence
- Pessimistic Concurrency Control
- Optimistic Concurrency Control
- Deadlocks and their detection/avoidance/prevention
- ACID Properties
- Nines Availability
- Active and Passive Replication
- Two-phase commit
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- Does an RPC always cross machine boundaries?
- Why is marshaling needed at all?
- What are conflicting operations and how can you use them to detect serial equivalence among transactions?
- Is locking a form of pessimistic or optimistic concurrency control?
- Does Google docs use pessimistic or optimistic concurrency control?
- What is one way to prevent deadlocks among transactions?
- What does “three nines availability” really mean?
- Why is Two-phase commit preferable over One-phase commit?
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Apply classical scheduling algorithms.
- Apply popular Hadoop scheduling algorithms.
- Know the internals of Apache Storm, a stream processing engine.
- Know how enormous graphs can be processed in clouds.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Classical Scheduling algorithms, including FIFO, Shortest Task First, and Round Robin
- Popular Hadoop schedulers including Capacity Scheduler and Fair Scheduler
- Internals of Apache Storm, a stream processing engine
- Internals of distributed graph processing engines, e.g., Google’s Pregel
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- Why is shortest-task-first optimal?
- What is the difference between the Capacity and Fair schedulers in Hadoop?
- What is a Storm topology?
- What is Gather-Apply-Scatter paradigm in distributed graph processing?
Readings and Resources
There aren't any readings required for this week, but you can look at the following documentation:week4:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know the internals of Distributed File Systems like NFS and AFS.
- Know the internals of Distributed Shared Memory systems.
- Know what’s inside a sensor mote and why networks of them are needed.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Distributed File Systems: Why they’re different from single-node file systems
- Internals of NFS
- Internals of AFS
- Distributed Shared Memory: How processes can share memory pages while communicating via messages
- Invalidate protocols in Distributed Shared Memory systems
- Sensor networks: Why they’ve emerged, what’s inside them, where they’re used, and what are the challenges
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- Why are Distributed File Systems stateless?
- How does NFS provide transparency?
- Why is whole file caching a reasonable approach in AFS?
- When is invalidate preferable over update in Distributed Shared memory systems?
- Why can’t embedded operating systems be used in sensor motes?
- What is the disadvantage of using a spanning tree in sensor network, for aggregation?
Readings and Resources
There are no readings required for this week, but you can look at the following documentation:week5:
Goals and Objectives
After you actively engage in the learning experiences in this module, you should be able to:- Know and design security policies.
- Know, apply, and design security mechanisms.
- Know the common properties of various types of networks.
- See case studies of multiple real datacenter outages, and learn lessons from them.
Key Phrases/Concepts
Keep your eyes open for the following key terms or phrases as you complete the readings and interact with the lectures. These topics will help you better understand the content in this module.- Security policies vs. mechanisms
- Security mechanisms: encryption, authentication, authorization
- Common characteristics among the various types of networks that surround us
- Why do datacenters suffer outages and what can we do about it?
Guiding Questions
Develop your answers to the following guiding questions while completing the readings and working on assignments throughout the week.- What is the difference between confidentiality, integrity, and availability?
- What is the difference between authentication, authorization, and auditing?
- What is a replay attack?
- What is the difference between public-private key cryptography and shared key cryptography?
- What is a small world network?
- What is the difference between a small network and a power law network?
- What is the most common cause of datacenter outages?
- What are three techniques to mitigate datacenter outages?
Readings and Resources
No readings are required for this week. But you can look at the following documentations:- Availability Digest Report
- AWS Outage Post-mortem
- Facebook Outage Post-mortem
- The Planet Outage Post-mortem
