Distributed Systems

Computer systems composed of multiple interconnected computers that communicate and coordinate their actions through message passing to achieve a common goal. Examples - Microservices, cloud-computing platforms, peer-to-peer networks, Blockchains and Distributed Databases.

Why distributed systems?

1. Scalable
   1. can handle large data volumes and number of transactions
2. Reliable/Resilient
   1. Reliability is a measure of the probability that a system will fail in a given period
   2. ensuring that the system doesn’t go down
   3. eliminate every single point of failure - i.e. being fault tolerant
      • this is done through redundancy/replication of both components and data
      • Replication means sharing information to ensure consistency between redundant resources, such as software or hardware components, to improve reliability, fault-tolerance, or accessibility
3. Available
   1. measure of the percentage of time that a system, service, or a machine remains operational under normal conditions
   2. i.e. no downtime, operational 100% of the time
   3. Reliability often implies Availability, but not necessarily other way around
4. Efficient
   1. Latency/ Response time: delay to obtain first item
   2. Throughput/ bandwidth: no of items delivered in a time unit
5. Manageability/Maintainability
   1. Ease of operation and maintenance
   2. Ease of diagnosis - observability through traces and logs
   3. Easy to update - code and infrastructure updates
6. Modularity and flexibility
   1. to an extent - though that is not the main goal.

See also Evaluation of Software Systems

Tradeoffs

CAP Theorem