January 31, 2016 / Dennis Holinka
Topic 2 – Application Architecture layer
This week's posts go over the Application Architecture layer, the various perspectives, and related reflections in the blog.
Post 3 - Service Oriented Application Architecture
The integration technology of n tier architectures have been evolving over the decade. However, I have been witness to the return of a point to point paradigm that has been recast into a SOA and Nexus of Forces approach because of some much needed standards -at least that is the way it appears to me after 10 years. It seems that the best architecture ever built has been the Internet with its universal resource locator and its ability to link and integrate any and every page to any other page that may be needed. The point to point integration was thought to be to tightly coupled and low cohesion in regards to its integration and it was replaced with the advent of EAI and ESB type technologies whose point was to mediate and broker the connections between 2 points so that the many different connections would not have to be rewired when changes to their interfaces and logic needed to be updated. By intermediating the interfaces, the broker would prevent tight coupling between service points - producer and consumer from being trapped dependently in change crisis. Therefore, the Service Oriented Architecture (SOA) emerged that made everything a service and the ESB/EAI mediation layer would broker all integrations in and out of the enterprise. The outside would be brokered by a Service Gateway and the inside by an Integration Platform.
But now, more than ever, this approach of using services and integration has become paramount in what amounts to an integration explosion of being able to link every call to any other service in a way that doesn't make the application hard to maintain. Therefore, we have reverted to standards that allow for mediation to happen by the endpoints themselves. That is, there is no need to use a central broker based technology to provide the location and service endpoint independence that was the downfall of point to point integrations. Rather, by building endpoints that have built in mediation technology build in, we can go back to using point to point without all of the pitfalls.
The technology that has emerged to support SOA and provide this independence is "the general adoption of the WS-* specifications (also called wizees, a set of open Web service standards that are part of the Web Services Architecture), the ESB will become obsolete, and the endpoints will be independent of any central entity responsible for the integration." When "systems and platforms support the WS-* specifications, the need for a central piece of middleware to supply the major integration tasks could disappear. The different services in such an architecture act more in a peer-to-peer mode, as opposed to the more client-server mode behavior we see in an ESB architecture. Some argue that SOA has the philosophy to expose the endpoints as equal and free members of the IT landscape. Each endpoint is visible and addressable. Endpoints can enforce security or transactions by providing a usage policy (WS-Policy) that contains corresponding assertions."
What does all of this amount to? To me, a return of point to point or expressed above - peer to peer. So what does that have to do with the future of application architecture? It may be that the Nexus of Forces architecture for application architecture needs to become cast as a peer to peer architecture between applications in the app-store integrating as that point between them or as the architecture of strategic micro-services that are orchestrated in a unified ecosystem of service architecture patterns. It is this architecture that we need to discover, document, and reuse in order to help enterprises compete on Amazon-like features and the growing consumer expectations of a smart user experience that is made up of cloud, data / analytics, mobility, and social event / stream integrations. No matter how we look at this new application architecture, it is definitely better than where we have been even if it is a remake or an innovation on the old paradigm.
The integration technology of n tier architectures have been evolving over the decade. However, I have been witness to the return of a point to point paradigm that has been recast into a SOA and Nexus of Forces approach because of some much needed standards -at least that is the way it appears to me after 10 years. It seems that the best architecture ever built has been the Internet with its universal resource locator and its ability to link and integrate any and every page to any other page that may be needed. The point to point integration was thought to be to tightly coupled and low cohesion in regards to its integration and it was replaced with the advent of EAI and ESB type technologies whose point was to mediate and broker the connections between 2 points so that the many different connections would not have to be rewired when changes to their interfaces and logic needed to be updated. By intermediating the interfaces, the broker would prevent tight coupling between service points - producer and consumer from being trapped dependently in change crisis. Therefore, the Service Oriented Architecture (SOA) emerged that made everything a service and the ESB/EAI mediation layer would broker all integrations in and out of the enterprise. The outside would be brokered by a Service Gateway and the inside by an Integration Platform.
Figure: Mediation by Integration Platform (e.g. virtual - EAI, ESB, ETL, SOA Gateway, etc)
But now, more than ever, this approach of using services and integration has become paramount in what amounts to an integration explosion of being able to link every call to any other service in a way that doesn't make the application hard to maintain. Therefore, we have reverted to standards that allow for mediation to happen by the endpoints themselves. That is, there is no need to use a central broker based technology to provide the location and service endpoint independence that was the downfall of point to point integrations. Rather, by building endpoints that have built in mediation technology build in, we can go back to using point to point without all of the pitfalls.
The technology that has emerged to support SOA and provide this independence is "the general adoption of the WS-* specifications (also called wizees, a set of open Web service standards that are part of the Web Services Architecture), the ESB will become obsolete, and the endpoints will be independent of any central entity responsible for the integration." When "systems and platforms support the WS-* specifications, the need for a central piece of middleware to supply the major integration tasks could disappear. The different services in such an architecture act more in a peer-to-peer mode, as opposed to the more client-server mode behavior we see in an ESB architecture. Some argue that SOA has the philosophy to expose the endpoints as equal and free members of the IT landscape. Each endpoint is visible and addressable. Endpoints can enforce security or transactions by providing a usage policy (WS-Policy) that contains corresponding assertions."
Figure: WS-* Standards
What does all of this amount to? To me, a return of point to point or expressed above - peer to peer. So what does that have to do with the future of application architecture? It may be that the Nexus of Forces architecture for application architecture needs to become cast as a peer to peer architecture between applications in the app-store integrating as that point between them or as the architecture of strategic micro-services that are orchestrated in a unified ecosystem of service architecture patterns. It is this architecture that we need to discover, document, and reuse in order to help enterprises compete on Amazon-like features and the growing consumer expectations of a smart user experience that is made up of cloud, data / analytics, mobility, and social event / stream integrations. No matter how we look at this new application architecture, it is definitely better than where we have been even if it is a remake or an innovation on the old paradigm.
I admire what you have done here. I like the part where you say you are doing this to give back but I would assume by all the comments that this is working for you as well. mi tarayıcı yukarı kaydırma kapatma
ReplyDelete