Are you ready to continue the journey to software-defined infrastructure? In an earlier post, I explored Two Key Stops along the Road to SDI: Automation and Orchestration. These stops are essential milestones in the trip to the ultimate destination: an SLA Managed data center.
In the SDI maturity model, the Automation and Orchestration stages feed into the SLA Managed stage, but the truth is they alone won’t get you there. To get to your final destination, your applications must be written to take full advantage of a cloud environment. Cloud-aware apps are like the vehicles on the road to the SLA Managed data center.
In more specific terms, cloud-aware apps know what they need to do to fully leverage the automation and orchestration capabilities of the SDI platform. They are written to allow for expansion and contraction automatically to maintain optimal levels of performance, availability, and efficiency. They understand that in a cloud environment, there are multiple routes to a database and multiple systems available to process data. They, in essence, do not worry about redundancy as the automation and orchestration will manage it in the environment.
This is quite unlike the conventional approach to apps. Most of today’s apps are tightly coupled with a particular database and a certain set of infrastructure resources. They require items such as session persistence and connection management. If any of the links break—for example, the app loses its connection to the database—the app goes down and IT admins go into fire-drill mode as they scramble to bring the app back online. Over the past 20 years, we have done our best to automate the fire drill.
In a metaphorical view, we’re talking about the difference between baseball and football. In baseball, things pretty much proceed in a linear and predictable manner. There are few moving parts—there’s one pitcher throwing to one batter—and aside from the occasional base-stealer you pretty much know where all the players are at all times. This is the way things work with the conventional app.
In a cloud environment, things are more football-like. The players are all over the place and the same play can unfold in very different ways. When a receiver runs the wrong route, the play doesn’t come to a stop. The quarterback simply looks for other receivers who are in position to make a play. The cloud-aware app functions like a quarterback who improvises to keep the ball moving down the field.
Here’s where things get harder. It’s not a trivial undertaking to make apps cloud-aware. In the case of legacy apps, the code has to be pretty much rewritten from top to bottom to build in cloud-awareness or the legacy part needs to be wrapped in services so that cloud aware can happen around the legacy portion. So we’re talking about a lot of heavy lifting for your software developers.
The good news is you don’t have to do all of this heavy lifting at once. We’re still quite some time away from the day of the SLA Managed data center. We have to first build the integrated orchestration platforms and automation toolsets that enable a software-defined approach to the data center. The key is to understand that this day is coming, and begin taking steps to make your apps cloud-aware.
Any new apps should be written to be cloud-aware. As for your legacy apps, you won’t be able to rewrite them all at once, so you’re going to need to identify the apps that are most likely to benefit from cloud awareness as you move to software-defined infrastructure or just wrap them in services.
The wrapped applications can help move many critical apps to a more cloud-like environment without rewriting a lot of code. But those apps won’t be able to benefit from all of the goodness of an SLA Managed data center. In an SLA Managed world, software-defined infrastructure and the apps work in tandem to deliver optimal performance with minimal downtime.
These gains are made possible by the ability of the orchestration platform to move workloads and supporting resources around on the fly to meet the policies you set for your applications. When demand spikes, the SDI environment grabs the resources the app needs to keep performance in line with the required service levels, even if that means bursting to a public cloud to gain additional processing power.
If this sounds like IT nirvana, you’ve got it. In the SLA Managed data center, application downtime will be rare, and unpredictable application performance will seem more like a problem from the past than a constant threat in the present. You’ll be able to breathe easier when unusually large crowds of holiday shoppers converge on a particular app, because you’ll know that the backend systems will take care of themselves.
So that’s the 30,000-foot view of the last stretches of the road to SDI. If you consider where we are today and where we need to travel, you can see that we are talking about a long road, and one that can have many unique twists and turns. The key is to think about how you’re going to get to SDI, identify the vehicles that will move you forward, and then begin your journey.
Find me @EdLGoldman, and share your thoughts and comments.