Microservices

Logical comparison of monolithic vs microservices. Source: Clover 2017.
Logical comparison of monolithic vs microservices. Source: Clover 2017.

Microservice Architecture describes an approach where an application is developed using a collection of loosely coupled services. Previously, applications were based on centralized multi-tier architecture. This worked well in the age of mainframes and desktops. But with cloud computing and mobile devices, backend must be available at all times for a wide range of devices. Bug fixes and features must be delivered quickly without downtime or deploying the entire application.

Microservices are independently deployable, communicating through web APIs or messaging queues to respond to incoming events. They work together to deliver various capabilities, such as user interface frontend, recommendation, logistics, billing, etc.

Microservices are commonly run inside containers. Containers simplify deployment of microservices but microservices can run even without containers.

Discussion

  • What are Microservices?
    What are microservices? Source: MuleSoft Videos 2018.

    A microservice is an autonomous independent service that encapsulates a business scenario. It contains the code and state. Usually a microservice even contains its own data store. This makes it independently versionable, scalable and deployable. A microservice is loosely coupled and interacts with other microservices through well-defined interfaces using protocols like http. They remain consistent and available in the presence of failure. A microservice is independently releasable. Each microservice can be scaled on its own without having to scale the entire app.

  • What are the types of Microservices?
    Stateless vs stateful microservices. Source: Microsoft Docs 2017.
    Stateless vs stateful microservices. Source: Microsoft Docs 2017.

    Broadly, there are two types of microservices:

    • Stateless: Has either no state or it can be retrieved from an external store (cache/database). It can be scaled out without affecting the state. There can be N instances. Examples: web frontends, protocol gateways, etc. A stateless service is not a cache or a database. It has frequently accessed metadata, no instance affinity and loss of a node is non-evident.
    • Stateful: Maintains a hard, authoritative state. For large hyper-scale application the state is kept close to compute. N consistent copies achieved through replication and local persistence. Example: database, documents, workflow, user profile, shopping cart, etc. A stateful service consists of databases and caches and the loss of a node is a notable event. It is sometimes a custom app that holds large amounts of data.

    As a variation, one author has identified three types: stateless (compute), persistence (storage), aggregation (choreography). Aggregation microservices depend on other microservices, thereby have network and disk I/O dependence.

  • What are the types of microservices in a layered architecture?
    Microservices layered architecture. Source: Indrasiri 2017.
    Microservices layered architecture. Source: Indrasiri 2017.

    When we look at microservices as a layered architecture, we can identify the following types:

    • Core/Atomic services: Fine-grained self-contained services. No external service dependencies. Mostly business logic. Often no network calls.
    • Composite/Integration Services: Business functionality composed from multiple core services. Include business logic and network calls. Implement routing, transformations, orchestration, resiliency and stability patterns. Often interface to legacy or proprietary systems.
    • API/Edge Services: A selected set of integration services and core services offered as managed APIs for consumers. Implement routing, API versioning, API security and throttling.
  • How are microservices different from APIs?
    Microservices vs APIs explained. Source: Clark 2021.

    APIs are not microservices and microservices are not the implementation of an API. API is an interface. A microservice is a component. The term "micro" refers to the component and not the granularity of the exposed interfaces. Microservices can be used to expose one or more APIs. However, not all microservice components expose APIs.

  • What's the relationship between Microservices Architecture and Service Oriented Architecture (SOA)?
    Comparing architectures of SOA and microservices. Source: Despodovski 2017.
    Comparing architectures of SOA and microservices. Source: Despodovski 2017.

    SOA and Microservice architecture relate to different scopes. While SOA relates to enterprise service exposure, microservice architecture relates to application architecture. Both try to achieve many of the same things (creation of business functions as isolated components) but at a different scale. They differ in maintainability, granularity, agility, etc. SOA is a very broad term and microservices is a subset of that usage. Netflix noted that microservices is "fine-grained SOA". Microservices have been recognized as "SOA done right".

    Some microservices principles are really different to SOA:

    • Reuse is not the goal: Reuse of common components is discouraged due to the dependencies it creates. Reuse by copy is preferred.
    • Synchronous is bad: Making synchronous calls such as API or web services creates real-time dependencies. Messaging is used whenever possible between microservices.
    • Service Discovery at run time: Components are assumed to be volatile. So it's often the clients’ responsibility to find and even load balance across instances.
    • Data Duplication is embraced: Techniques such as event sourcing result in multiple independent "views" of the data, ensuring that microservices are truly decoupled.
  • What are the important principles to keep in mind while designing a Microservices Architecture?

    Broadly speaking the following principles are good to know while designing a microservices architecture: Modelling around a Business Domain, Culture of Automation, Hide Implementation Details, Highly Observable, Decentralise all things, Isolate Failure, Consumer First, Deploy Independently.

  • What are the advantages of using a microservices architecture?

    Advantages span both development and operations. Briefly we note the following advantages: Build and operate service at scale, Improved resource utilisation to reduce costs, Fault isolation, Continuous Innovation, Small Focused Teams, Use of any language or framework.

    Scalability comes because of the modularity that microservices enable. Because of containers, microservices have become easier deploy in various environments. Because of the isolation of services, there's also a security advantage: an attack on one service will not affect others.

    Microservices are about code and development. Containers are about deployment. Containers enable microservices. Containers offer isolated environments, thus making them an ideal choice for deploying microservices. However, it's possible to deploy microservices without containers. For example, microservices can be deployed independently on Amazon EC2. Each microservice can be in its own auto-scaling group and scaled independently.

  • Why should I probably not adopt microservices for my app?
    Challenges with microservices. Source: MacVittie 2018.
    Challenges with microservices. Source: MacVittie 2018.

    Since application is distributed across microservices, such distributed systems are harder to manage and monitor. Operational complexity increases as the number microservices and their instances increases, particularly when they are dynamically created and destroyed. Network calls can be slow and even fail at times. Being distributed, maintaining strong consistency is hard and application has to ensure eventual consistency.

    Microservices require more time for planning and partitioning the app. They should be designed with failure in mind. When you are building a Minimum Viable Product (MVP) or experimenting to assess what works or can add value to business, then a monolithic approach is faster and easier. Adopt microservices only when you need to scale after your idea and its business value is proven.

    If you're managing a legacy application, the effort to migrate to microservices might involve considerable cost. The technology stack may be considerably larger than a monolithic app. Applications have a great dependence on networking and its performance. Microservices can be tested independently but to test the entire application is more difficult.

Milestones

1970

Although not directly influencing the birth of microservices, it's been noted that from an architectural/philosophical perspective similar ideas existed in the 1970s: Carl Hewitt's Actor Model; or pipes in UNIX that connect many small programs, each of which did something specific rather than having a large complex program that did many things.

1990

This is the decade when Service-Oriented Architecture (SOA) starts to gain wider adoption. The term itself is first described by Gartner in 1996. The idea is to use services as isolated and independent building blocks that collectively make up an application. Services are loosely bound together by a standard communications framework.

2000

To enable applications to communicate with one another over a network, W3C standardizes SOAP. Over the next few years, WSDL and UDDI gets standardized as well. XML is the preferred format of communication. Through this decade, Web Services, a web-based implementation of SOA, becomes popular over proprietary methods such as CORBA or DCOM.

2009

Netflix redefines the way it develops apps and manages operations by relying completely on APIs. This later develops into what we today call microservices.

May
2011

The term microservices gets discussed at a workshop of software architects near Venice, to describe a common architectural style that several of them were exploring at the time.

2012

REST and JSON become the de facto standard to consume backend data. These would later prove to be foundational for inter-service communication when building microservices-based apps.

Mar
2013

Closely related to microservices, Docker, a computer program that performs operating-system-level virtualization (containerization), gets open sourced.

Jun
2014

Kubernetes, a container orchestration system for automating deployment, scaling and managing containerized applications, gets open sourced by Google.

Nov
2015
Survey shows microservices entering mainstream adoption. Source: NGINX 2015.
Survey shows microservices entering mainstream adoption. Source: NGINX 2015.

An app developer survey from NGINX shows that microservices are entering mainstream with about two-thirds either using or investigating them. Adoption is higher with medium and small companies.

References

  1. Amundsen, Mike. 2017. "Three Types of Microservice Components." RESTFest 2017, September 16. Accessed 2018-11-11.
  2. Avram, Abel. 2013. "Docker: Automated and Consistent Software Deployments." InfoQ, March 27. Accessed 2018-10-13.
  3. Bertram, Adam. 2017. "7 reasons to switch to microservices — and 5 reasons you might not succeed." CIO, June 20. Accessed 2018-11-11.
  4. Chotin, Matt. 2018. "Containers and Microservices: Two Peas in a DevOps Pod." The New Stack, March 16. Accessed 2019-08-05.
  5. Clark, Kim. 2017. "Microservices vs SOA." IBM Developer, YouTube, July 6. Accessed 2018-10-12.
  6. Clark, Kim. 2021. "Microservices vs APIs." IBM Support and Training, YouTube, February 16. Accessed 2021-03-28.
  7. Clover, Joey. 2017. "Microservices are hard — an invaluable guide to microservices." HackerNoon, October 19. Accessed 2018-10-17.
  8. Despodovski, Rade. 2017. "Microservices vs. SOA – Is There Any Difference at All?" DZone, November 12. Accessed 2018-11-04.
  9. Erl, Thomas. 2004. "Introduction to Web Services Technologies: SOA, SOAP, WSDL and UDDI." InformIT, Pearson, September 03. Accessed 2018-11-11.
  10. Evans, Josh. 2017. "Mastering Chaos - A Netflix Guide to Microservices." InfoQ, YouTube, February 22. Accessed 2018-10-14.
  11. Fowler, Martin. 2014. "Microservices." GOTO2014 Berlin, GOTO Conferences, YouTube, January 15. Accessed 2018-10-14.
  12. Fowler, Martin. 2015. "Microservice Trade-Offs." July 01. Accessed 2018-11-11.
  13. Hausenblas, Michael. 2015. "The Origins of Microservices." Microservices Practitioner Articles, Medium, November 16. Accessed 2018-11-11.
  14. Hummel, Guy. 2018. "Microservices Architecture: Advantages and Drawbacks." Blog, CloudAcademy, October 18. Accessed 2018-11-11.
  15. Indrasiri, Kasun. 2017. "Microservices Layered Architecture." Microservices in Practice, Medium, September 15. Accessed 2018-11-11.
  16. Lewis, James and Martin Fowler. 2014. "Microservices." March 10. Updated 2014-03-25. Accessed 2018-11-11.
  17. MacVittie, Lori. 2018. " Mainstream Microservices Mania: Challenges Increasing with Adoption." F5 Networks, July 02. Accessed 2018-11-11.
  18. Metz, Cade. 2014. "Google Open Sources Its Secret Weapon in Cloud Computing." Wired, October 6. Accessed 2018-10-13.
  19. Microsoft Docs. 2017. "Orchestrating microservices and multicontainer applications for high scalability and availability." .NET, Microsoft Docs, May 19. Accessed 2018-11-11.
  20. MuleSoft. 2018. "What are microservices?" MuleSoft Videos, YouTube, January 24. Accessed 2018-11-04.
  21. NGINX. 2015. "The Future of Application Development and Delivery Is Now." NGINX. Accessed 2018-11-11.
  22. Natis, Yefim. 2003. "Service-Oriented Architecture Scenario." Gartner, April 16. Accessed 2018-11-11.
  23. Newman, Sam. 2015. "Principles of Microservices." Devoxx, YouTube, November 12. Accessed 2018-10-13.
  24. Osowski, Rick. 2018. "Introduction to microservices." IBM Developer, November 10. Accessed 2018-11-11.
  25. Photon Infotech. 2016. "A Brief History of APIs & Microservices [Infographic]." Digital Transformation Insights, July 14. Accessed 2018-11-11.
  26. Posta, Christian. 2017. "About When Not to Do Microservices." Blog, September 28. Accessed 2018-11-11.
  27. Roeser, Robert B. 2018. "Four Myths about Microservices." Netifi, via Medium, July 25. Accessed 2019-08-05.
  28. Scholl, Boris. 2015. "Introduction to Microservices." Microsoft Visual Studio, YouTube, November 18. Accessed 2018-10-12.
  29. Tozzi, Christopher. 2018. "6 Reasons Not to Adopt Microservices." Container Journal, January 19. Accessed 2018-11-11.
  30. Weiss, Todd R. 2018. "How WeatherBug Uses Microservices Without Containers." The New Stack, July 23. Accessed 2019-08-05.
  31. Young, Greg. 2016. "The Long Sad History of MicroServices." IT Arena, via SlideShare, October 28. Accessed 2018-11-11.

Further Reading

  1. Lewis, James and Martin Fowler. 2014. "Microservices." March 10. Updated 2014-03-25. Accessed 2018-11-11.
  2. Evans, Josh. 2016. "Mastering Chaos - A Netflix Guide to Microservices." QCon San Francisco 2016, November 08, via InfoQ, YouTube, 2017-02-22. Accessed 2018-11-11.
  3. Tech Primers. 2017. "How to Design Microservices Architecture? Uber Architecture - A Case Study." Tech Primers, YouTube, July 31. Accessed 2018-11-11.
  4. Bertram, Adam. 2017. "7 reasons to switch to microservices — and 5 reasons you might not succeed." CIO, June 20. Accessed 2018-11-11.

Article Stats

Author-wise Stats for Article Edits

Author
No. of Edits
No. of Chats
DevCoins
Avatar of user arvindpdmn arvindpdmn
7
1
1239
User avatar arjun
3
0
947
DevCoins due to articles, chats, their likes and article hits are included.
1461
Words
2
Authors
10
Edits
3
Chats
8
Likes
16K
Hits

Cite As

Devopedia. 2021. "Microservices." Version 10, March 28. Accessed 2024-06-26. https://devopedia.org/microservices
Contributed by
2 authors

Last updated on
2021-03-28 13:54:07

Improve this article
design system microservices cloud

See Also

Article Warnings

  • Readability score of this article is below 50 (45.1). Use shorter sentences. Use simpler words.