Topic #4 - Infrastructure Architecture

Does Cloud Spell the Death of Infrastructure Architecture?

Selby (2016) explores the idea how the role of the infrastructure architect has changed or will exist with the move towards the cloud. The debate is short lived on whether or not the role is dead but he certainly points out that is will/has changed. I would agree with his point the concept of infrastructure is changing however for us it is slowly changing. We have a significant investment in our data center and will continue to have that investment. One of the major hurdles we have with moving towards the cloud is that our data center is considered and asset since we own the equipment in the data center. The idea of cloud challenges our capitalization policy that we currently have in place for IT assets. That is one area that is a heated discussion when it is brought up. However, that aside I think there is a real need to figure out the right balance of cloud vs on premise infrastructure that is needed to fulfill business needs. Selby (2016) starts to touch on this with the following statement, "the Infrastructure Architect is already fast becoming the Integration Architect (IA) - a specialist in tying together services the business requires from the physical connectivity right up to the presentation layer and application delivery - and doing so on the most appropriate platform, be it in-house or with an array of cloud providers." I would certainly agree with that sediment. I think as architects we need to figure out the right blend of infrastructure it going to be key. Like Selby (2016) says, the business demands, "consumer IT ‘just works’ and seamlessly integrates, so why can’t my corporate application." It seems that this is where the role of infrastructure architecture is headed, however for us it will be a while before we get all the way there because we will still need to focus on our on premise infrastructure as well.

Reference
Selby, Rob (2016). Does Cloud Spell the Death of Infrastructure Architecture. Retrieved October 5th, 2016 from http://www.adapt.com/whats-new/blog/does-cloud-mean-the-death-of-infrastructure-architecture.html

Data center infrastructure architecture redrawn to meet new IT demands

In December of 2013, Courtemanche and Boisvert (2014) interviewed Henrique Cecci research director at Gartner Inc. about cloud and data centers at the Gartner Data Center Conference. Cecci mentions that application deployment is driving the need for new approaches to infrastructure. He also mentions the need to be able to deploy infrastructure faster than ever before in order to keep up with the application demands.

For my organization we made a significant investment in virtualization several years ago and now are reaping the benefits of that decision. There are a couple different aspects that virtualization has allowed us to excel in. The first is provisioning, we can provision a set of servers to support a new deployment in the matter of days (maybe hours) if needed. We do not need to wait on procurement of hardware to enable these new applications. Second is disaster recovery, we have been able to establish a disaster recovery strategy with complete failover to a secondary data center. I suppose maybe even a 3rd benefit is our backup strategy, we are able to take snapshots of servers prior to upgrades and revert back to the original if an upgrade goes south.

I also find the part of the Courtemanche and Boisvert (2014) interview where Cecci talks about modernizing the data center to meet demands. For me, I always looked at hardware refreshes as a way to get the most modern processing and storage power available. I hadn't thought about the efficiencies that are gained along with that. Total cost of ownership analysis like this, "If you are reducing the temperature in the data center by 1 degree Celcius, that may represent 4% energy costs in the data center", are something that doesn't seem like much on the surface but apply that thinking to infrastructure architecture as a whole and there might be more cost justification in something like a hardware refresh.

Reference
Courtemanche, Meredith and Boisvert, Michelle (2014). Data center infrastructure architecture redrawn to meet new IT demands. Retrieved October 5th, 2016 from http://searchitoperations.techtarget.com/feature/Data-center-infrastructure-architecture-redrawn-to-meet-new-IT-demands

Scaling Microservices Architectures in the Cloud

I found the article from Saini (2015) a nice short review of architecture when looking at micro services. This has been something I have been thinking about for the last few months and how we might leverage this idea of micro services. To gain the full advantage of micro services you want to be able to scale your architecture to process additional load as needed. As Saini (2015) points out you cannot accomplish this with the traditional work horse server instead you need to be able to use a multi-cluster setup. For me this is a challenge and the biggest part of it is around licensing. When I look at our licensing around our middleware servers they are core based and that becomes a problem if/when you want to scale rapidly. This is a challenge I'm trying to get my head around when it comes to hosting micro services. We are also currently mixing some batch processing, services and messaging. These three integration patterns have different infrastructure requirements, this is another area I'm evaluating to see better understand what we can/should do from an infrastructure perspective.

Reference
Saini, Atul (2015), Scaling Microservices Architectures in the Cloud. Retrieved October 9th, 2016 from https://www.fiorano.com/blog/infrastructure-architecture/