Data Center Modernization Using a Hybrid Cloud Strategy


Today’s data centers face challenges for efficiency and performance that can be met using cloud technologies. However, determining the best cloud model and which workloads can securely be located where, or even which cloud service provider (CSP) would deliver the best return on investment (ROI), can be difficult for many organizations.

Research by IDC projects that 90+ percent of enterprises around the world are relying on a mix of on-premises, or dedicated, private clouds as well as several public clouds and legacy platforms to meet their infrastructure needs. Gartner predicts that, by 2025, 80 percent of enterprises will be using entirely cloud-based infrastructures. And, according to research by Flexera, 80 percent of survey respondents are taking a hybrid approach that combines using both public and private clouds. Altogether, cloud technologies are fast becoming a staple of enterprise computing.

Most often, there is either a hybrid cloud or multicloud strategy in place, the difference being that a hybrid cloud infrastructure blends two or more types of clouds (e.g., private, public, on-premises), while a multicloud, which primarily combines public cloud infrastructures, blends multiple clouds of the same type. For the purposes of this article, we will focus on a hybrid cloud approach.

Challenges Facing Data Centers

Enterprises need to maintain data security, ensure infrastructure performance, and make the most of legacy infrastructure for ROI. This also means they require the agility to change direction and scale based on dynamic usage needs, which also means future-proofing their investments—including for cloud technologies.

A foundational enablement for considering a hybrid cloud strategy is to see the public cloud as an extension of organizational infrastructure. This becomes important, as building additional on-premises data center capacity is expensive and time-consuming, and it cannot scale down once the investment has been made. A public cloud can easily scale for bursts in usage and enable applications at the edge of the network, meaning closer to remote users, with nominal up-front cost.

That said, decisions need to be made about which workloads and applications can (or should) run in the public cloud. Your organization might have older versions of operating systems that are not supported by CSPs. In that case, rearchitecting applications can be an additional project that takes time and resources.

Organizations also need to avoid vendor lock-in, to support remote and distributed employees with on-demand access to data, process data in real time, and to increase cloud computing capacity without increasing the data center. These challenges can be mitigated with a hybrid cloud environment. At the same time, enterprises must have a cloud strategy in place to avoid over- or under provisioning for cost containment.

What to Know about Hybrid Clouds

A hybrid cloud strategy enables agility in meeting business needs, from data privacy to avoiding vendor lock-in for any one CSP. Additionally, a hybrid cloud can scale to meet dynamic business needs, including peak activity, as workloads can be interchanged as demand and costs shift.

This can be critical for such applications as autonomous driving, industrial robots, and AI algorithms that need near-real-time data processing, which benefit from a hybrid cloud strategy to exploit the speed of on-premises data centers and the large data storage capacity of the cloud.

Another benefit of hybrid cloud is that you can have discrete applications or even cloud-native applications to meet your business needs. There is no need to hold back on cloud-native applications, but you also do not need to power legacy applications into the cloud until you are ready. The hybrid cloud gives you the flexibility to move at a comfortable pace as needed.

In the new “work from anywhere” model, organizations with hybrid cloud data centers can leverage their infrastructure so it is accessible to their remote employees. This can enhance performance, efficiency, and automation of both tasks and workforce productivity.

By leveraging automation in the cloud for configuring and provisioning resources, managing applications, and handling routine tasks, your organization can be more efficient and maximize IT staff for projects that require a higher level of skill.

A hybrid cloud also means reduced spend on infrastructure as fewer servers are needed on-premises.

To discover whether a hybrid cloud strategy could make sense in meeting business needs, enterprises should ask a few questions, as follows.

  • What is our current infrastructure? Is it scalable?
  • What is our current solution stack?
  • What workloads can be processed more efficiently by using the cloud?
  • Have we optimized our systems?
  • What is our density in terms of storage needs, now and projected?

The answers to these questions can help qualify the optimal cloud strategy for your organization.

Cloud Technologies for Security and Performance

When used in a hybrid cloud strategy, data centers can lower spend on infrastructure, gain manageability through APIs, and have consistent visibility of data whether at rest or in motion. In all cases, the data must be protected.

Securing your resources in the public cloud is always a shared responsibility between your organization and your chosen CSP. That said, top-tier CSPs invest heavily in making their data centers as secure and reliable as possible because they host workloads for hundreds of thousands of customers. Your CSP acts as an extra security layer between your processor and software.

Cloud technologies can provide an array of effective defensive measures to protect data and minimize potential attack surfaces. Intel-powered cloud technologies have built-in hardware-level security, such as with Intel® Software Guard Extensions (Intel® SGX), which isolates specific application code and data in memory so it is inaccessible from the outside.

It is important to ensure you have portability for flexibility in shifting workloads to different clouds as your needs change over time. Your chosen hybrid strategy should deploy to where the capabilities exist to optimize for the needs of given workloads.

You can also choose to optimize your on-premises environment so it mirrors the performance optimizations of cloud technologies by using Intel® Field Programmable Gate Arrays (Intel® FPGAs) to support crypto acceleration for enhanced hardware performance and Intel® Optane™ persistent memory for memory tiering. As a resource in monitoring different performance metrics, Intel Telemetry Collector (ITC) can collect references, including power and thermal statistics, performance counters, process activities, and threads and operating system–level disk, network, and memory statistics.

Your organization should determine what type of workloads should be distributed in the various environments of a hybrid cloud approach. This decision also prevents future “cloud sprawl” and avoids overprovisioning. A common miscalculation is that when an application is seen as highly valuable, it can be assigned greater cloud bandwidth when the reality is it may not need significant cloud bandwidth. Or when a newer cloud technology is perceived as more expensive, it may actually save resources because the uptime is more efficient, meaning shorter, so it is more cost-effective.

As examples of what workloads should be distributed where, consider the usage patterns from within your existing environments. You may want to retain sensitive or highly confidential data, especially if in a highly regulated industry in an on-premises data center. Other types of workloads to consider are virtual machines (VMs), business software, databases and backups, disaster recovery, development, analytics, and collaboration tools.

Future-proofing Your Cloud Investment

Changes to any cloud infrastructure in the future are inevitable, as needs, usage, and costs are dynamic. Too often, enterprises do not think about the potential changes that can make future migrations challenging. The triggers for these changes can be anything from cost increases to security risks to company policies to the need to repatriate (meaning to pull back into an on-premises environment). Preparing for such changes in advance makes them easier in the long run and maximizes your investments.

A hybrid cloud easily accommodates when some workloads need to stay on-premises and others need to shift based on demand. The good news with a hybrid cloud solution architecture is it helps avoid vendor lock-in, as cloud reliance is shared among CSPs and data centers. And by choosing Intel-based x86 architecture, which is common, future portability is simplified, as optimizations and configurations are built into the system. Optimizations in the software stack use the features of that hardware, so when migrating data to a new cloud, there are no performance issues. There is no need to rearchitect your infrastructure. And when you use containers to migrate, the process is even more streamlined.

Intel offers Node Feature Discovery (NFD), which is a Kubernetes add-on that helps pick the right node to run containers and facilitate intelligent scheduling of workloads, depending on need and platform capabilities. This orchestration of system resources and capabilities helps deliver streamlined application performance.

Another option is to use microservices, which are lightweight, loosely coupled services that can be used to build cloud-based applications. A microservices architecture is agile, scalable, and supports rapid deployment, dynamically operating multiple instances on a single or multiple servers to support workload demands. Individual microservices are often containerized to improve portability and scalability.

This kind of portability, when dictated by risk, performance, and cost, is what makes a hybrid cloud future-proof. By understanding your current infrastructure, solution stack, workloads, operating system, kernel level, usage, optimizations, configurations, and density for storage, you have a solid foundation for making responsible decisions around your cloud infrastructure.

If your organization is not yet convinced that committing data and applications to any cloud solution is the right strategy for your data center, a hybrid cloud could be the answer. Hybrid cloud eases security concerns by allowing you to keep critical resources in on-premises data centers and move other workloads to the public cloud.

Connection’s suite of cloud services is backed by a team who can help organizations navigate the complexities of various cloud environments. Connection’s team of experienced frontline experts are standing by to tailor a hybrid cloud strategy for your data center needs. To begin, just ask for a Cloud Landscape Optimization (CLO) assessment.

Intel® technologies may require enabled hardware, software, or service activation.
No product or component can be absolutely secure.
Your costs and results may vary. © Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.