By Jeff Schultz
Many organizations are migrating their applications and infrastructure into the cloud. That’s why 80 percent of all IT budgets will be committed to cloud solutions in the near future.
There are some great benefits to operating in the cloud versus operating with on-premises software. With the cloud, you’ll have a recurring operating expense—instead of a big capital expenditure.
The cloud can also make capacity planning easier. At least, the cloud makes it easier to add or remove capacity when needed. If you need to add more capacity to a system tomorrow, you’ll have the flexibility to spin up more capacity.
But, there are some drawbacks to moving into the cloud, too. You can lose sight of exactly how much capacity you have—and how much you’re using. Peter Drucker’s regularly quoted saying rings true for capacity too—“If you can’t measure it, you can’t manage it.”
In practice, doing capacity planning for the cloud isn’t as easy as it seems. To do it right, there are eight aspects you need to consider.
1. Service Level Agreements
Your organization has service level agreements (SLAs) that you need to meet.
The primary goal of most SLAs is to avoid downtime and ensure continuous service delivery. This is typically done through ensuring availability and creating a back-up and recovery plan in the event of downtime. This requirement comes from the business.
When you’re doing capacity planning for the cloud, you still need to consider your SLAs.
2. Utilization Patterns
Utilization patterns tell you how utilized your systems, servers, and applications are.
Knowing your utilization patterns is crucial to capacity planning—and it will be important in the cloud as well. This is how you’ll know what is expected at what time. After all, your utilization may vary by time of day or by time of year.
Monitoring application utilization helps you properly manage capacity. Consider this scenario. You have 4 vCPUs and 2 gb on-premises at 10 percent utilization. But in the cloud, you may only have 1 vCPU and 500 mg of memory. Capacity planning can help you manage that transition.
3. Workload Analytics
Workload analytics is an assessment of your workloads today. Doing workload analytics is an important step for moving your workloads into the cloud.
You’ll need to consider why workloads change—and what happens when they do. It will be important to review your historical data and identify trends. You should also take into account what the business is projecting—and how that will impact your future workloads. For example, planning to add a new customer next quarter will affect capacity demands.
From there, you can make a smart plan for moving your workloads into the cloud.
4. What to Put in the Cloud
One of the biggest things to consider when moving into the cloud is what to put in it. While you eventually might migrate everything to the cloud, you’ll probably start out with just a few applications, systems, and servers.
So, you need to figure out what makes sense to move to the cloud right off the bat. Applications that require a large memory requirement may not be good cloud candidates without modifications.
There are application dependencies you might consider, particularly if an application you want to move to the cloud is dependent on an application you’ll be keeping on-premises.
You’ll also need to consider availability. Moving an application into the public cloud can give you greater availability. But if you move an application into the private cloud, it gives you more flexibility and control.
As with everything else in IT these days, security and compliance are important. You’ll want to make sure you’re making smart moves to lock down your infrastructure as you migrate into the cloud.
You should also consider support. What impact will moving certain applications into the cloud make on support?
Finally, are there any external dependencies you need to consider for your applications? An example might be an application dependent on a confidential data source or dependent on another cloud application where availability is a concern.
5. Data Management Policies
Data management policies ensure that your organization is properly managing and retaining data. These policies will remain important as you migrate into the cloud.
Some key areas to pay attention to include:
You’ll need to make sure that when you move an application into the cloud, you’ll still comply with your data management policies.
6. Proper Cloud Instance Sizing
There’s a common mistake that many organizations make when it comes to capacity planning in the cloud. And that’s not properly sizing cloud instances.
Proper cloud instance sizing ensures that you have the right instances based on your utilization. (This is where understanding your utilization patterns is important). While cloud vendors make recommendations, it’s ultimately up to you to make sure you’ve sized your cloud instances for your needs.
As you size your cloud instances, you should make sure you understand how much customization is possible. Will you be able to size up or down easily? How much will the flexibility to do so? Finally, you should consider how many pre-defined sizes make sense right now.
7. Disaster Recovery
Having a disaster recovery plan is critical to every IT organization. That won’t change when you move into the cloud.
You’ll still need a disaster recovery plan. By creating it, you should be able to answer What is the recovery time? for your applications, systems, and servers. You should also be able to determine what the potential impact—and cost—downtime due to disaster would have on your business. If the recovery window is four hours, then you can build a DR plan accordingly.
8. Business Requirements
Today, the business counts on IT as a strategic partner. Their business requirements are your IT requirements.
So, when moving to the cloud, you need to map business needs to capacity requirements. This might mean understanding that the business will grow 30 percent in three months—and translating that growth into infrastructure needs.
Your goal should be to support business objectives. And you should be able to tell the business—in their terms (cost, response times, availability, etc.)—what’s going to happen in the next three to six months.
Do Capacity Planning in the Cloud
Learn more about going into the cloud. Watch our recent webinar, How to Do Capacity Management in the Cloud.
About the author: Jeff Schultz is a product specialist for capacity management at HelpSystems. He’s answered questions for customers and ‘tire kickers’ for more than 20 years. Jeff boasts a background in helping IT organizations get up to speed on their IT environment for their capacity planning and performance management objectives.