Where should your workloads live?
Should you rent, lease, or buy your next residence? To make a decision, you weigh a multitude of variables, talk to people you trust, and then make a decision that works for your personal situation.
Similar logic applies to workload placement decisions in a Hybrid IT environment. There are many deployment choices to consider and no one-size-fits-all approach, because different workloads have different characteristics and each business is unique. You need to weigh the key decision factors, compare different options, and then match the workload with the appropriate destination based on specific business objectives.
In a recent Frost & Sullivan survey, 51 percent of IT decision-makers cited “assessing the optimal deployment model for workloads” as a difficult challenge in implementing their Hybrid IT environment. It’s complicated, and the landscape changes constantly. Not too long ago, traditional IT was the default platform. Here, the application, service, and workload were all nicely tied together. The terrain was familiar and mostly contained within your data center.
Workload placement patterns are shifting. According to a recent 451 Research study, 60 percent of workloads still reside in traditional or non-cloud infrastructure, compared with 40 percent for cloud workload deployments. The authors predict that "this mix will change dramatically in two years, essentially reversing the workload figures to 58 percent of workloads deployed in the cloud and 42 percent deployed on non-cloud infrastructure.”
Keep in mind that not all clouds are public. Cloud is about delivering the agility and speed the business expects. If you adopt a Hybrid IT strategy, your environment will consist of different deployment environments, including traditional IT, on-prem private cloud, managed cloud, software as a service (SaaS), and public cloud. Each has pros and cons. It’s a good idea to consider the full range of deployment options for each workload placement decision you make.
Traditional IT consists of computing resources that reside within your data center and are configured to meet your particular needs.
Like traditional IT, private clouds are customized for each customer deployment and are part of your data center. Using automation and orchestration software integrated with hardware, developers and end users can access the services they need through a self-service catalog. Once a service is selected, resources are automatically and rapidly provisioned to meet workload and application development needs.
With public cloud, customers consume shared, scalable services that are delivered over the Internet.
Managed cloud services are a hosted model that provides the benefits of the cloud with access to services and tools to support the unique needs of each business. Unlike public cloud, managed cloud services are managed by a third-party resource.
SaaS is a model designed to support the needs of multiple customers. It includes shared, hosted software services like Salesforce and Office 365, delivered over the web and licensed by end users.
Hybrid IT typically involves a mix of environments and many variables. You'll need to consider four main variables when making a workload placement decision.
1) Hybrid IT security and compliance. Each day brings new exposure and attack surfaces, increasingly sophisticated cyberattacks, and complex regulatory pressures.
When considering off-prem environments, you need to know exactly what your vendor will do if there is a problem and understand up front what it would take to bring the workload back on-premises if needed.
How much risk can you tolerate? What will it take to keep your enterprise secure and compliant? These are complex calculations that vary by geography and company. In the U.S., for example, you might determine that a certain workload is appropriate to run on a public cloud. In Europe, you might opt to run the same workload in an on-prem cloud environment, due to European data sovereignty rules and data sensitivity concerns in your business.
2) Total cost of ownership. Comparing TCO across platforms can feel like comparing apples and puppies. It's fairly easy to calculate TCO for a group of dedicated applications across a given set of servers, switches, and storage. Even then, it can be hard to compute actual cost per application. Cost for public cloud is often calculated on a price per minute for each virtual machine (VM)/minute, but those prices can be volatile. With managed cloud, the price is calculated on a VM/hour basis, with a set price for each contract. And there are many cost factors to consider with private cloud. They include hardware, software, data center costs, business continuity needs, and personnel costs, among others.
Don’t assume that public cloud is always the cheapest option. Let's say you plan to build an enterprise-grade private cloud or invest in software-defined infrastructure to speed application delivery. When you go to present your investment business case, be prepared to demonstrate that the TCO is either less than, equal to, or greater than the public cloud. And don’t focus on cost alone. Figure out the price for each option, and be ready to differentiate them in business value terms. Take the time to define an apples-to-apples comparison that you can use to inform your decision.
3) Application performance. According to Frost & Sullivan, “60 percent of IT decision-makers cite concern about application performance as a reason not to move a workload to public cloud.” Public cloud is accessed over the Internet, and the infrastructure is shared. Virtual machines compete for the same resources, introducing latency potential. But who determines resource prioritization in a shared environment? Consider whether your workload requires low-latency performance. There are workloads, such as high-frequency trading in the financial services industry, where application performance needs to support high-speed transactions and delays are unacceptable. Determine what you consider an acceptable delay and whether your workload can tolerate inconsistent latency. Depending on your application performance needs, a private cloud may deliver the performance you ultimately require.
4) Control is a factor that cuts across almost all of the other variables. The right level of control varies with company culture and policies. IT seeks to stay relevant by providing business users what they need, when they need it, at the right economics. You need to provision the right resources quickly to match the needs of each workload. You need to manage the workload from end to end, and you may need to manage it centrally. According to Frost & Sullivan, 80 percent of IT decision-makers cite "loss of control over application" as a reason to avoid deploying a workload in the public cloud.
Which of your workloads distinguishes you from the competition? For these workloads, how much control do you or your line-of-business counterparts feel you need? Would you be comfortable running them with a third party? Just because you can put a workload in the public cloud doesn’t mean you will or should.
Different workloads have different needs. 451 Research shows both current and projected deployment trends broken down by workload type, including email and collaborative, web and media, data and analytics, application development, business applications, and shared IT workloads. I encourage you to take a look at this paper if you’re interested in the specific trends occurring by workload type or want to learn more about how to make an informed decision.
Also, data center technology has evolved rapidly in recent years. There are ways to get the speed and agility of public cloud for a variety of workloads using traditional IT. Software-defined options such as composable and hyperconverged infrastructure with cloud management software can rapidly deliver traditional and cloud-native applications. You should also consider on-demand services that combine the agility and economics of public cloud with the security and performance of on-prem IT.
Where to place your workloads: Lessons for leaders
- Just because you can place your workload on the cloud doesn't mean you should.
- Carefully consider regulatory issues that can impact your workload placement.
- Comparing TCO between the various solutions can be very difficult. Make sure your comparison points are accurate.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.