Given this trend, IT leaders must become adept in determining how to allocate their resources between their data centers and public and private cloud options. Also, they must manage performance proactively wherever their technology is located.
So how should you manage this increasingly complex environment?
Should You Go Public or Private?
Once you’ve decided to migrate certain applications to the cloud or house new projects there, one of the first questions is whether to use a public cloud service provider or to create a private cloud.
The private cloud offers some of the same benefits as public cloud offerings. You’ll enjoy the elasticity to meet the demands of your internal and external customers and the ability to meter and chargeback expenses to cost centers within your organization. Also, many still feel that the private option offers increased security.
Private clouds are, however, not without challenges. You still have to manage the associated hardware, hypervisor, and virtualization. You will also continue to incur capital expenses. So consider whether your company has the skills to support a private cloud and decide whether you want to keep absorbing substantial capital expenses.
With a public cloud provider, you’ll also gain elasticity and an easy way to allocate costs between departments. But you won’t have to manage and maintain hardware, software, and facilities. You can also rest assured that the top cloud service providers hire the best people and are always developing new services that are designed to increase the efficiency and effectiveness of the technology that supports your business.
One of the greatest difficulties occurs before you move. You cannot simply make a forklift move of all the technology that supports the applications you’re migrating. For the greatest efficiency, performance and to take advantage of cost center billing, you’ll require a new infrastructure architecture.
To help in creating one, you must understand the workload and performance of your servers in your brick-and-mortar data center. If you have a database that’s doing 500 IOPS a second with a service time of less than a millisecond, you want to ensure the cloud environment provides equivalent in performance at a minimum.
Monitoring a Hybrid Cloud Environment
To re-architect your IT environment, you need to monitor what’s happening in your data center today as well as your performance while testing a cloud provider and over the long term. The data that a monitoring solution provides is a prerequisite for planning and managing any IT environment.
While all cloud providers offer their own monitoring systems, they have their limitations. On the plus side, they provide useful data on CPU, memory, and billing. Cloud providers monitoring solutions, however, typically do not go deep enough on disk information. For example, you likely will not be able to get insights into metrics on queue depth and block size. Since these impact throughput and the user experience, you will need these parameters make decisions on whether to go bare metal or virtual.
In a virtual environment, servers are treated like livestock, not pets. They’re transient. Data moves from one server to the next as demand dictates. Because of the lack of permanency, cloud service vendors keep a minimal amount of historical data.
But even if historical data is not relevant to your cloud service provider, it is important to you.
You want to know that despite the constantly changing environment, the service you test and sign up for is what you receive on an ongoing basis. You need to have data that backs up performance you received when you initially tested the cloud solution. That’s what you should have specified in your Service Level Agreement (SLA). Should the performance you receive degrade later, you have proof of the service level decline that you can offer to your vendor.
Lastly, whatever monitoring solution your cloud service providers offer, it will only provide data on the cloud portion of your infrastructure. Instead, you need a monitoring solution that provides a holistic picture of your hybrid environment — your data center and the cloud. This bird’s eye view enables you to make strategic decisions about your hybrid cloud infrastructure.