Arch2Arch Tab BEA.com
Syndicate this blog (XML)

Going beyond the Silo: 5 Steps to Federation

Bookmark Blog Post

del.icio.us del.icio.us
Digg Digg
DZone DZone
Furl Furl
Reddit Reddit

Dain Hansen's Blog | April 21, 2008   1:08 PM | Comments (0)


You’ve heard the buzz word of Federation as a silver-bullet to just about everything. In this blog we’ll show you 5 important steps to getting your IT federated and some key architecture patterns to consider as a way to leverage specific Federation practices to help you out of that silo’d thinking.

Step_0Step 0: Still Stuck in the Silo: You’ve got multiple projects in your enterprise and possibly more than one location for a data-center. Where do you keep your services? Everywhere and no where. You’ve got services for billing, for orders, different systems for payments, and finally fragmented views of customer information. To top it all off, you have no standardized way to implement processes that call these services. Today this is one of the most common patterns that we’ve seen where the Enterprise Service Bus (ESB) acts like an integration hub within a project silo.

Silo-centric ESBs lead you to these 3 key headaches:

  • Domain or Project ESBs don’t provide enterprise-wide visibility of services and greatly complicate how you control and manage it
  • More time is spent on integrating the integration, as opposed to improving re-use.
  • Flexibility and reuse is limited, greatly reducing the benefits of SOA

So let’s fix it!

Step_1Step 1: Big Bus, Little Bus.  Most confuse federation with simply connecting ESBs together. But one thing that is important is you can separate one of the ESBs and make it simply used for generic pass-through routing to the other project-centric multiple ESBs. The other ESBs can provide additional service mediation which may include: transformation, validation, message enrichment, or even orchestration. The advantage of building out the deployment model this way is that the ESB at the top has visibility of the routes and acts a control mechanism for the subservient ESBs. You should make sure that the inter-domain routes are guaranteed by putting these messages on JMS/SAF (Store and Forward) or WS-RM. In some cases you may not have the same vendor for each of the ESBs – but this doesn’t limit your ability to connect between them using standard messaging protocols. It also is a good reason to leverage the top-level bus to give better consistent visibility of end-to-end routes.

AquaLogic Service Bus supports reliable JMS/SAF, WS-RM connections or even local transport (for highly optimized localized deployments). In addition, AquaLogic Service Bus can be scaled to support linear scalability as more and more projects and services are provisioned.

Step_2Step 2: Federated Data Access. Consistent and reliable data access that spans multiple data sources/systems has been a pressing challenge that many enterprises struggle with. When solved correctly, it allows for single-source-of-truth of data which can be used for broader information. For example, if you wanted all the information about your current customers today, how many systems would you have to access? In this model we need to ensure that the data is federated how it is accessed so that it can be re-used as a single service. That service can be passed into the ESB to provide greater re-use across the organization. Because the data sources stem from multiple project locations it is extremely important to ensure that the data is consistent, accurate, and up-to-date.

AquaLogic Data Services Platform supports Federated data access has optimized connectivity for AquaLogic Service Bus and is the backbone of Information as a Service (IaaS) architecture.

Step_3Step 3: Enterprise Wide Business Processes. Remember that your business doesn’t run in a silo, even if your IT is managed that way. Business processes often go across lines of business and may break down the barriers between organizational projects. If we’re only thinking of BPM as human-centric processes, we don’t have any challenges here for federation; unfortunately most business processes in organizations actually do touch automated system-centric processes. These integrations between the Business-centric and the IT-centric collide. For example, billing processes need to often access a host of CRM and financial systems as well as an approval workflow process. To solve this both BPM and ESB need to be integrated together to provide service-enabled business processes to span this federated model. These integrations need to support reliability guarantees and be integrated into the design-time components of the two solutions.

AquaLogic Service Bus and AquaLogic BPM today support tight integrations (both design-time and run-time) for better managing broad federation of business processes that leverage highly federated services underneath.

Step_4Step 4: Management and Security of Composite Services. The complexity of federation is no different than the complexity of composite services across an organization. We need to be able to consistently manage how services behave as they are messaged between ESBs, Data Services, BPM, and other service applications. We need to be managing the Service Level Agreements (SLAs) and be able to enforce their explicit behavior as a policy. For example, when data is extracted from a system, it might need to be scrubbed for privacy considerations – let’s say we want to hide someone’s full social security number when a call center is verifying the last four digits. These types of fine grained entitlement policies can help to better secure services and provide governance on the run-time behavior.

AquaLogic SOA Management provides management of services beyond simply an ESB, but across projects and across domains. Together with AquaLogic Enterprise Security can provide fine-grained entitlements for greater control of data and information.

Step_5Step 5: SOA Governance for better control and visibility of Federated deployments. The original objective of Federation was so that you can get better control and visibility of the deployments. Because services don’t always go through a single ESB or even a single big bus ESB, there is an even greater level of abstraction that is required. In this case the Service Component Architecture (SCA) can help. By leveraging a Repository/Registry that supports SCA, you can ensure that composite deployments can be visualized, managed and deployed across multiple projects and domains. For example, one team develops inventory services, another team develops billing services; consequently a third team can leverage the two to create new services around order management. Because these services are located in a single repository/registry, scorecards about the performance characteristics, dependencies on who is calling them, can lead to faster-time-to market of bringing these new services into production and ensuring that they continue to be leveraged effectively.

AquaLogic Registry Repository helps better control federated integration implementations of ESB, Data Services, BPM and provide out-of-the-box capability for SCA for better managing composite services.

What are some of the steps in Federation you are still struggling with? Feel free to send over your comments!

References

 

 

 

 

 

 


Comments

Comments are listed in date ascending order (oldest first) | Post Comment



Only logged in users may post comments. Login Here.

Powered by
Movable Type 3.31