carloscastilla - Fotolia
Hyper-converged infrastructure can help address many of the challenges that come with implementing today's workloads on a traditional three-tier architecture. Hyper-converged platforms are easier to deploy, manage and scale than typical IT infrastructures, freeing up IT resources for other projects. However, organizations implementing HCI must take into account a number of factors when preparing to migrate workloads to that platform. Here, we look at five important considerations that IT should keep in mind before making the move to HCI.
Make security your top priority
HCI systems typically employ virtualization and software-defined technologies to deliver application services. As part of this process, hyper-converged technology abstracts the underlying compute, storage and network resources, resulting in a shared set of services with a common security profile. As a result, a single root attack can compromise the entire system, disrupting services and exposing sensitive data. The longer an attack goes undetected, the more severe the damage.
At the same time, HCI is a multilayered architecture with discrete control, data and management planes. Legacy approaches to data center security were not designed for HCI and are often not enough to meet the challenges that come with such an environment. You need data security that can protect resources at both the physical and logical layers, working across the three planes. Before moving your workloads to hyper-converged technology, you should have a security strategy in place that can accommodate the underlying infrastructure and its evolving and variable workloads.
To this end, the tools you use to protect your workloads should be designed with HCI security in mind. At the same time, you should take advantage of the technologies built into the HCI to help protect the environment. For example, VMware vSAN provides software-defined encryption, while Microsoft Hyper-V supports shielded VMs. Also, be sure to adhere to security best practices, such as using policy-based security, virtual networks, access control lists and platform-wide encryption.
Understand your workloads
Before you move ahead with hyper-converged technology, fully assess the workloads you want to implement within the HCI environment. To this end, you must understand what types of applications you'll be running, the amount of data they'll be processing and the nature of that data. You should also note the technologies that drive the applications, such as whether they use containerization, microservices or infrastructure as code.
Application types can vary widely, so you need to take a complete inventory of their requirements. For example, you might want to run offerings that support big data analytics, AI, virtual desktops or IoT. Each has specific performance, storage and availability requirements. You should understand exactly what those requirements are, considering such issues as the number of users, when they work and how the applications will need to scale going forward.
Another factor to consider is whether all application components will be deployed to the HCI environment. Some applications, for example, might need to access data stored on a cloud platform or in an external database, in which case the components running within the hyper-converged environment must be able to integrate seamlessly with those systems outside it.
Evaluate the HCI platform
Whether you plan to purchase a new hyper-converged system or use one you've already deployed, you must thoroughly evaluate its capabilities to ensure it can support your expected hyper-converged workloads. Your goal is to learn as much about the system as possible. What are the performance limitations? Can you run multiple hypervisors? Does it support containers or microservices? What is the storage capacity? How much of that capacity is usable space?
As part of your assessment, be sure to consider issues related to availability and scalability. In the early days of hyper-converged technology, compute and storage resources had to scale together, which often meant overprovisioning one or the other. This has begun to change, however, making HCI more flexible. But you still must understand existing limitations and exactly how you scale each system. You should also determine whether you can use commodity hardware when it comes time to replace or add components.
When you're evaluating an HCI system, also examine your application requirements. For example, if you have specific network connectivity needs, be sure the platform can support the necessary protocols and topologies and it includes the physical ports needed to make those connections. Or, if you'll be running applications that have high GPU requirements, make certain the hyper-converged environment can meet those needs.
Plan your management strategy
Once you've deployed your hyper-converged workloads to an HCI environment, you'll need to manage the applications and the data that drive them. Long before you do that, however, you should be thinking about how you'll carry out that management. You must be able to deploy, update and patch your applications as efficiently as you can with your current infrastructure.
Start by verifying whether the management tools you already use will extend into the HCI environment. Legacy management tools might have a difficult time seeing and accessing all the components in a system that relies so heavily on virtualization and software-defined architectures. You might need to upgrade your existing tools or purchase new ones. You might also need to update any customizations you have in place that extend your current tools.
Also, take into account how you plan to back up and archive data, as well as implement DR. In addition, you should consider how you will monitor your applications and data and the HCI components that support them. Although hyper-converged technology is based on the premise of simplifying IT operations, the added abstraction layers can make it more difficult to troubleshoot issues if something goes wrong, another reason an effective monitoring strategy is so critical.
Evaluate supporting systems
When evaluating a hyper-converged system, you should assess how it will integrate with third-party platforms and other data center systems. We mentioned earlier that your applications might need to access data stored on a cloud platform or external database, but there are other systems to consider, such as identity management, security, content services, orchestration tools and system configuration tools, among numerous others.
You need to fully understand the implications of deploying the HCI system to your existing data center environment, assessing legacy infrastructure components, connectivity to external environments and the network on which the HCI system will be deployed. The better you understand the HCI platform's ability to integrate with its surrounding environment, the better you can prepare for moving your workloads to that platform.