Chapter 14. TCI - Key-Value Pair Store

Question 1

What function does the Key-Value Pair Storage provide? Why is it important? How can be communicated with it?

Answer 1

• provides the functionality to store or retrieve the value of a key
• most Key-Value stores provide REST-APIs to support GET/PUT/DELETE which help with operations over HTTP

Question 2

Why is the Key-Value Pair Storage in Clouds important?

Answer 2

• its important, because in a distributed and dynamically-scalable environment we need a single-point of truth endpoint
• for example we want each node to look up specific configuration values before performing an operation
• for that all nodes can reach out to a central location and retrieve the value of a desired variable or key

Question 3

What are examples of Key-Value Pair storages?

Answer 3

• etcd
• Consul KV
• Zookeper

Question 4

What is etcd?

Answer 4

• open source distributed key-value pair storage
• uses Raft consensus algorithm for communication between instances of distributed multi-instance scenarios

Question 5

What is consensus?

Answer 5

• fundamental problem in fault-tolerant distributed systems
• involves multiple servers agreeing on values
• once they reach a decision on a value, that decision is final
• typical consensus algorithms make progress when any majority of the servers is available
• if more than half the servers fail they stop making progress (but will never return an incorrect result)

Question 6

Name features of etcd

Answer 6

• can run standalone or in a distributed cluster
• in cluser mode, for high availability, can gracefully handle leader election during network partitions and tolerate machine failures, including the leader
• allows users or services to watch the value of a key, to then perform certain operations as a result of any change in the particular value

Question 7

For what is etcd used?

Answer 7

• store connections, configurations, cluster bootstrapping keys and other settings
• service discovery in conjunction with tools like skyDNS
• metadata and configuration data for service discovery
• container management

Question 8

What are benefits of etcd?

Answer 8

• open source key-value pair storage
• default for Kubernetes container orchestrator
• provides reliable data storage across cluster of machines
• fast, benchmarked at 10.000 writes/s per instance
• provides seamless cluster managmeent across a distributed system
• secured by optional SSL client certificates for authentication and offers very good documentation

Question 9

What is Zookeeper?

Answer 9

• open source project of the Apache Software Foundation
• centralized service for mainting configuration information, providing distributed synchronization together with group services for distributed applications

Question 10

What does Zookeeper aim to do and implement?

Answer 10

• aims to provide a simple interface to a centralized coordination service that is also distributed and highly reliable
• implements consensus, group management and presence protocols on behalf of applications

Question 11

What are Zookeepers use-cases?

Answer 11

• implement node coordination in a distributed environment
• manage cloud node memberships while coordinating distributed jobs
• as a backing store for distributed and scalable data structures
• election of a high availability leader
• light-weight failover and load balancing manager