Why is Capacity Management Important?
In today’s world of constant and ever accelerating change we have the tendency to discard the old and look for the new. This is true for all sorts of consumer items, but it is often true for the ideas we have and the processes we use too. New must be better just because it’s new, so anything that sounds antiquated starts off with a low rating. Capacity management is a bit like that.
It sounds like something we did twenty or thirty years ago (it was!). And, it sounds like it’s all about physical bits and pieces, not the clouds and containers of today’s virtual landscapes. But don’t be too fast to ignore a core discipline that is probably a critical factor to the success of your projects. What is your capacity for success?
The truth about infinite resources
As alternatives to on-premise infrastructure, we have gone from “your-mess-for-less” outsourcing to the promise that you needn’t even worry about compute power and storage if you use the cloud. The fact is that someone has to pay for whatever capacity is used, and that will be you.
The ability to pay as you go is great when you don’t use very much, not so great when consumption starts to grow. Even worse, your consumption might be untracked and growing continually. The real truth is that you need capacity management tools even more when your resources are so fluid and provided on-demand. The truth is that you might just find that your business services have an infinite capacity to use more!
Read our eBook
Controlling Cloud Costs with Capacity Management
Capacity management has been used for decades to optimize on-premise resources. Now, as cloud environments transform IT, it is being extended to enable holistic planning, management, and optimization of all your resources in one place and at the same time. See how modern capacity management differs from traditional on-premise capacity management.
It’s about the services, stupid
Another thing we have been doing in IT for 30 years is talking about aligning IT with lines of business (LOBs). This is the way that modern capacity management is practiced and is complimentary to cutting edge practices.
In all enterprises of any size there will be key business services that are cross-platform and possibly leverage hybrid cloud architectures. These complex alignments of infrastructure — and the transactions that pass through them — need to be thought of as the single entity that they really are from a business and end-user perspective.
Key to your success is to know the capacity of these interconnected pieces to sustain transaction growth. If pieces are shared the problem becomes even harder to manage. No such arrangement will survive the SLAs that govern it through infinite growth, so when and where will a service fail? That is the question that capacity management, practiced well, will answer.
Capacity issues are best addressed at design time, or migration time if you are planning on moving a service to the cloud. What capacity will you need to go live with successfully in either case? With the DevOps revolution, design time is not a point of origin, but a regular event in cycles that are routinely weekly updates, and much shorter in extreme DevOps driven environments.
And, if I have lost my mainframe and IBM i audience here, it’s not that different for them. Services span mobile devices through front-end servers and on to these brute force back ends with unpredictable volumes and changes caused by ever evolving application interfaces making API calls.
In all cases, state-of-the-art IT involves capacity management as an integral part of DevOps practices. You measure seamless through the Ops-Dev-Ops transitions and the same tools both detect bugs during development and alert of problems in production.
Real time would be real useful
Traditionally, capacity management has been inherently based on historical data. Data is collected, aggregated to intervals and analyzed to project historical trends. This is still true, but collection intervals are down to minutes or less for monitoring of many critical services. This is real time by some definitions. But when it comes to automated operations, real time means essentially “as fast as you can,” and too slow is too late; fail.
Sometimes, we do capacity management to plan server upgrades weeks and months ahead. But in today’s online-always world, provisioning of new virtual resources is often just-in-time to allow services to scale to demand. Real-time capacity management is now the requirement. I would differentiate this from automation driven by APM solutions by the depth and sophistication of the analytics and the blending of historical data with real-time awareness.
Success by design
Your capacity for success is largely determined by your understanding of the problem, your knowledge of the possible solutions, and your ability to design to your needs.
Syncsort Capacity Management is not your grandfather’s capacity management. Take a second look if you don’t use these disciplines to manage your environments yet.
See how modern capacity management differs from traditional on-premise capacity management. Read our eBook: Controlling Cloud Costs with Capacity Management