Impact of Restful Web Architecture on Performance and Scalability

Impact of Restful Web Architecture on Performance and Scalability

• 1.
  Impact of RESTful web architecture on performance and scalability Sanchit Gera 2B Computer Engineering University of Waterloo
• 2.
  Overview • Background – Web Services –Performance – Scalability • Introduction to REST • Characteristics of RESTful design • Summary • Conclusion 2
• 3.
  Web Services • Means of exposing functionality or data • Analogous to web pages for machines •Facilitate integration between applications • Tend to be flexible and client agnostic CLIENT SERVICE Request Response 3
• 4.
  Performance • Driven by a number of factors, some of which are impacted by architectural decisions • Performance discussed in terms of –Relates to issues such as network latency and limited network bandwidth – Minimizing data-trips between server and client – Minimizing “message overheads” 4
• 5.
  Scalability • Need to support large number of concurrent instances or interactions • Scaling up- increasing the capacity of services, consumers, and network devices (vertical scaling) • Scaling out - distributing load across services and programs (horizontal scaling) 5
• 6.
  Introduction to REST • REpresentationalState Transfer • An architectural style for web services •Open ended specifications • Utilizes the inherent design of the Web for scalability and performance 6
• 7.
  Characteristics of REST • Uniform interface • Stateless system •Client-server architecture • Cacheable • Layered System • Code on Demand (Optional) 7
• 8.
  Stateless System • Server does not store any context between requests • Each request should be independent and self containing –Self descriptive messages • Any session state is held on the client • Imposes restrictions on the communication allowed between services and consumers 8
• 9.
  Stateless System 9
• 10.
  Stateless System • Frees up service memory resources, allowing the service to scale with the number of concurrent requests • Less complicated to design and distribute across servers (horizontal scaling) •Reduces network performance by increasing transmission of repetitive data 10
• 11.
  Cacheable • Service consumers can cache and reuse response message data • Responses from the service to consumers are labeled as cacheable or non-cacheable •Improves runtime efficiency and allows for “lazy replication” • Boosts system performance by reducing user perceived latency 11
• 12.
  Cacheable • Caches work mostly, but not exclusively with GET requests • Implemented either implicitly, explicitly or through negotiation •Better user perceived performance • Lowers perceived network latency • Reduces load on the database boosting scalability 12
• 13.
  Layered System • Solution can have multiple intermediary architectural layers • No one layer can see past another •Request-response mechanism must be transparent to end client • Every layer has a single high level purpose, boosting separation of concerns • Allows us to add features like a gateway, a load balancer, or a firewall to accommodate system scaling. 13
• 14.
  Conclusion • Brief overview of REST and its principles • Analysis of REST principles based on scalability and performance –Statelessness : Performance (Negative), Scalability (Positive) – Cacheable : Performance (Positive), Scalability (Positive) – Layered System : Performance (Negative), Scalability (Positive) 14