Skip to content
mythz edited this page Apr 11, 2013 · 9 revisions

Why ServiceStack?

Developed in the modern age, Service Stack provides an alternate, cleaner POCO-driven way of creating web services.

  • Simplicity
  • Speed
  • Best practices
  • Model-driven, code-first, friction-free development
  • No XML config, no code gen - convention-driven!
  • Smart. Infers intelligence from strongly typed DTOs
  • .Net and Mono
  • Highly testable - services are completely decoupled from HTTP
  • ServiceStack has been around for 4+ years

Features:

  • XML, JSON, SOAP and more!
  • Built-in validation - with smart fluent syntax!
  • Automatic Exception handling
  • Caching (Memcached and Redis are supported out-of-the box!)
  • Authentication/authorization (with built-in OAuth providers for Twitter/Facebook)

Define web services following Martin Fowlers Data Transfer Object Pattern with ServiceStack


Service Stack was heavily influenced by Martin Fowlers Data Transfer Object Pattern:

When you're working with a remote interface, such as Remote Facade (388), each call to it is expensive. As a result you need to reduce the number of calls, and that means that you need to transfer more data with each call. One way to do this is to use lots of parameters. However, this is often awkward to program - indeed, it's often impossible with languages such as Java that return only a single value.

The solution is to create a Data Transfer Object that can hold all the data for the call. It needs to be serializable to go across the connection. Usually an assembler is used on the server side to transfer data between the DTO and any domain objects.

The Request and Response DTO's used to define web services in ServiceStack are standard POCO's while the implementation just needs to inherit from a testable and dependency-free IService<TRequestDto>. As a bonus for keeping your DTO's in a separate dependency-free .dll, you're able to re-use them in your C#/.NET clients providing a strongly-typed API without any code-gen what-so-ever. Also your DTO's define everything Service Stack does not pollute your web services with any additional custom artefacts or markup.

Service Stack re-uses the custom artefacts above and with zero-config and without imposing any extra burden on the developer adds discover-ability and provides hosting of your web service on a number of different physical end-points which as of today includes: XML (+REST), JSON (+REST), JSV (+REST) and SOAP 1.1 / SOAP 1.2.

WCF the anti-DTO Web Services Framework

Unfortunately this best-practices convention is effectively discouraged by Microsoft's WCF SOAP Web Services framework as they encourage you to develop API-specific RPC method calls by mandating the use method signatures to define your web services API. This results in less re-usable, more client-specific APIs that encourages more remote method calls.

Unhappy with this perceived anit-pattern in WCF, ServiceStack was born providing a Web Sevice framework that embraces best-practices for calling remote services, using config-free, convention-based DTO's.

ServiceStack encourages development of message-style, re-usable and batch-full web services

Entire POCO types are used to define the request and response DTO's to promote the creation well-defined coarse-grained web services. Message-based interfaces are best-practices when dealing with out-of-process calls as they can batch more work using less network calls and are ultimately more re-usable as the same operation can be called using different calling semantics. This is in stark contrast to WCF's Operation or Service contracts which encourage RPC-style, application-specific web services by using method signatures to define each operation.

As it stands in general-purpose computing today, there is nothing more expensive you can do than a remote network call. Although easier for the newbie developer, by using methods to define web service operations, WCF is promoting bad-practices by encouraging them to design and treat web-service calls like normal function calls even though they are millions of times slower. Especially at the app-server tier, nothing hurts performance and scalability of your client and server than multiple dependent and synchronous web service calls.

Batch-full, message-based web services are ideally suited in development of SOA services as they result in fewer, richer and more re-usable web services that need to be maintained. RPC-style services normally manifest themselves from a client perspective that is the result of the requirements of a single applications data access scenario. Single applications come and go over time while your data and services are poised to hang around for the longer term. Ideally you want to think about the definition of your web service from a services and data perspective and how you can expose your data so it is more re-usable by a number of your clients.

Difference between an RPC-chatty and message-based API

public interface IWcfCustomerService
{
    Customer GetCustomerById(int id);
    List<Customer> GetCustomerByIds(int[] id);
    Customer GetCustomerByUserName(string userName);
    List<Customer> GetCustomerByUserNames(string[] userNames);
    Customer GetCustomerByEmail(string email);
    List<Customer> GetCustomerByEmails(string[] emails);
}

contrast with an equivalent message based service:

public class Customers : IReturn<List<Customer>> 
{
   public int[] Ids { get; set; }
   public string[] UserNames { get; set; }
   public string[] Emails { get; set; }
}

Any combination of the above can be fulfilled by 1 remote call, by the same single web service - i.e what ServiceStack encourages!

Fewer and more batch-full services require less maintenance and promote the development of more re-usable and efficient services. In addition, message APIs are much more resilient to changes as you're able to safely add more functionality or return more data without breaking or needing to re-gen existing clients. Message-based APIs also lend them better for cached, asynchronous, deferred, proxied and reliable execution with the use of brokers and proxies.

Comparatively there is almost no win for a remote RPC API, except to maybe hide a remote service even exists by making a remote call look like a method call even though they're millions of times slower, leading new developers to develop inefficient, brittle systems from the start.

Community Resources



  1. Getting Started
    1. Create your first webservice
    2. Your first webservice explained
    3. ServiceStack's new API Design
    4. Designing a REST-ful service with ServiceStack
    5. Example Projects Overview
  2. Reference
    1. Order of Operations
    2. The IoC container
    3. Metadata page
    4. Rest, SOAP & default endpoints
    5. SOAP support
    6. Routing
    7. Service return types
    8. Customize HTTP Responses
    9. Plugins
    10. Validation
    11. Error Handling
    12. Security
  3. Clients
    1. Overview
    2. C# client
    3. Silverlight client
    4. JavaScript client
    5. Dart Client
    6. MQ Clients
  4. Formats
    1. Overview
    2. JSON/JSV and XML
    3. ServiceStack's new HTML5 Report Format
    4. ServiceStack's new CSV Format
    5. MessagePack Format
    6. ProtoBuf Format
  5. View Engines 4. Razor & Markdown Razor
    1. Markdown Razor
  6. Hosts
    1. IIS
    2. Self-hosting
    3. Mono
  7. Advanced
    1. Configuration options
    2. Access HTTP specific features in services
    3. Logging
    4. Serialization/deserialization
    5. Request/response filters
    6. Filter attributes
    7. Concurrency Model
    8. Built-in caching options
    9. Built-in profiling
    10. Messaging and Redis
    11. Form Hijacking Prevention
    12. Auto-Mapping
    13. HTTP Utils
    14. Virtual File System
    15. Config API
    16. Physical Project Structure
    17. Modularizing Services
  8. Plugins
    1. Sessions
    2. Authentication/authorization
    3. Request logger
    4. Swagger API
  9. Tests
    1. Testing
    2. HowTo write unit/integration tests
  10. Other Languages
    1. FSharp
    2. VB.NET
  11. Use Cases
    1. Single Page Apps
    2. Azure
    3. Logging
    4. Bundling and Minification
    5. NHibernate
  12. Performance
    1. Real world performance
  13. How To
    1. Sending stream to ServiceStack
    2. Setting UserAgent in ServiceStack JsonServiceClient
    3. ServiceStack adding to allowed file extensions
    4. Default web service page how to
  14. Future
    1. Roadmap
Clone this wiki locally