Click here to Skip to main content
Click here to Skip to main content
Go to top

Creating a managed service factory

, 21 Aug 2007
Rate this:
Please Sign up or sign in to vote.
Host any number of services, regardless of whether they are remote or local within a "Server" construct and access that server easily.

Introduction

I had a plan. I wanted to create a way of accessing services in multiple locations from any location. Each location could have one or more services which may be duplications or different. That is a really abstract way of thinking about it, but I eventually came up with a solution.

After using Reflector to view the Team Foundation Server source code, which gave me many pointers for which I would like to thank the TFS Team at Microsoft, I came up with the following solution:

image

This allows you to host any number of services, regardless of whether they are remote or local within a "Server" construct and to access that server easily. Remote and local services can exist side by side in the same server, for example, you could have an ActiveDirectoryClientService hosted in a ClientServer.

From the server level (IClientServer), you can request a service based on its type or on an interface that it implements. To achieve this, the IClientService interface has a property called contracts that loops through all of the interfaces that the client service object implements and pulls out only those interfaces that have the ClientServiceContractAttribute added to them. Thus, you need to make sure that all of your interfaces (e.g., ITeamServer) have that attribute regardless of whether it also has a ClientServiceContractAttribute.

From the factory level, you can again request either services by type or by implementation, but this request is cross server. If you have two servers added to the factory and they both implement different contracts, you can pull any one back you like. If both servers have the same implementation, you can choose whether you bring back the first or a random implementation. The potential for a load balancing solution is there, but not implemented.

Below is a diagram of my TFSEventHandlerServer that includes both local and remote services. For ease of development and because it is a common thing to do, I created the WcfClientServiceBase class that provides a number of features that allow less code to be written in the individual service client classes.

image

In this example, you can call any of the ClientServices with the same code:

Get the server with a call:

Dim Server As Servers.TFSEventHandlerServer
Server = Servers.ClientServerFactory.GetServer(Of Servers.TFSEventHandlerServer)_
         (New System.Uri(http://server:6873))

And get client services with the Subscriptions being a local service and TeamServers as a remote one.

Dim Subscriptions As Clients.SubscriptionsService
Subscriptions = Server.GetService(Of Clients.SubscriptionsService)()
Dim TeamServers As Clients.TeamServersService
TeamServers = Server.GetService(Of Clients.TeamServersService)()

As I hope, you can see this makes it easier to implement many features with an enterprise environment. All you have to know is what services are available where. There is also the possibility that a lookup service could be implemented that would allow the Factory to bring you services when you do not even know there they are!

I hope this helps those trying to find a way to achieve the same goals without being too restrictive. You can download the code from my CodePlex TFS Event Handler project.

License

This article, along with any associated source code and files, is licensed under The Microsoft Public License (Ms-PL)

Share

About the Author

Martin Hinshelwood
Instructor / Trainer naked ALM
United States United States

About the Author: Martin has worked with many customers in government, finance, manufacturing, health and technology to help them improve their processes and deliver more. He provides management and technical consulting that intends to expose processes and practices to gain transparency, uncover impediments to value delivery and reduce cycle-time as part of an organisations path to agility. Martin is a Professional Scrum Trainer as well as a Visual Studio ALM MVP and Visual Studio ALM Ranger. He writes regularly on http://nakedalm.com/blog, and speaks often on Scrum, good practices and Visual Studio ALM.

You can get in touch with Martin through naked ALM.

Follow on   Twitter   Google+

Comments and Discussions

 
-- There are no messages in this forum --
| Advertise | Privacy | Mobile
Web02 | 2.8.140916.1 | Last Updated 21 Aug 2007
Article Copyright 2007 by Martin Hinshelwood
Everything else Copyright © CodeProject, 1999-2014
Terms of Service
Layout: fixed | fluid