The Current State of Servers: 2020 Trends and Expectations
In this article
According to an IDC study done in 2019, global public cloud services spending is expected to more than double by 2023. The biggest portion of this increase will be Software as a Service (SaaS), with Infrastructure as a Service (IaaS) coming in second. It is anticipated that industries such as professional services, discrete manufacturing, and banking will represent the largest share of spending in public cloud services.
This begs the question: "Are companies moving away from physical servers and migrating their environments to public cloud services entirely?"
What we're seeing and anticipate continuing to see in 2020 is that most organizations are opting to create hybrid environments that utilize the best of both worlds. Organizations are using a combination of physical servers and public cloud services, motivated largely by balancing cost and performance.
It is safe to say that physical servers are still an important part of the data center in 2020. What we have learned from our customers is that they do not want to run everything in the public cloud. Instead, they prefer to consume many of the public cloud services that are costly or complicated for them to build and operate on-premises.
Some of these types of services include databases, large-scale object stores, scalable and reliable web services, analytics and artificial intelligence (AI) services. The need for these services is real, especially among developers, and is often why customers start using public cloud services in the first place.
Today, the vast majority of workloads for most customers still exist on-premises and for legitimate reasons: cost, control, security policies, data gravity, workforce skill set and business process integration. Moving these on-premises workloads to public cloud service providers often does not yield any significant business benefits other than allowing them to leverage cloud services like the ones mentioned above.
As previously stated, we are seeing organizations leverage some combination of both on-premises infrastructure and public cloud services. There are a couple of different ways the industry is addressing this need.
First, there are those companies that are moving aspects of their on-premises data centers to the public cloud. This includes on-premises infrastructure running in or near the cloud using VMware Cloud (VMC) on Amazon Web Services (AWS), Microsoft Azure VMware Solution, Google Cloud VMware Engine or co-location (colo) with low-latency cloud connectivity.
Secondly, for those organizations moving from public cloud to on-premises infrastructure, public cloud "stacks" are being constructed. Organizations are building these stacks using platforms/services such as Microsoft Azure Stack, AWS Outposts or Google Anthos.
Organizations continue to place importance on many of the same server specifications that they have in the past. Performance improvement is especially important in today's world as the amount of data we are accessing, analyzing and storing grows.
Despite wanting better and increased performance from their infrastructure, companies also want a reduction in their physical footprint. Power efficiency is crucial in reducing an organization's footprint. As always, cost is a driving factor in server needs and requirements.
In addition, customers are increasingly focusing on the management of their infrastructure. Organizations want and need to be able to access their data quickly. To do this, effective and efficient infrastructure management is critical. Storage capabilities to support software-defined storage (SDS) and hyper-converged infrastructure (HCI) use cases are also becoming an important component of what customers are looking for in server technology.
It's also important that today's servers can support acceleration technologies, such as persistent memory, GPUs, FPGAs and ASICs, as well as remote bare metal provisioning, software/firmware upgrades, hardware upgrades/expansions and automated and standardized software stack deployments.
We continue to see the gradual move away from legacy mainframe and Unix systems toward x86 architectures. There is still little to no movement in ARM-based systems outside of a select few service providers.
Looking at it through the optics of the form factors, we continue to see a slight shift from blade servers to modular and rack mount systems. This is largely to support more storage and/or acceleration technologies. HCI and SDS are driving the need for storage density, while high performance computing (HPC), analytics and AI are driving the need for acceleration technologies.
In the short-term, we expect to see more enhancements around storage density coupled with greater performance. With advanced network acceleration technologies and protocol built in and around the server, it will in turn benefit from faster processing. Additionally, developments in centralized management and monitoring of physical servers (i.e. infrastructure management) will create more efficiency in the data center.
Looking farther out on the horizon, disaggregated systems will help solve many of the scaling and cost challenges that come with current blade and rack mount form factors. Intel Rack Design (RSD) architecture offers flexible and efficient use of infrastructure in the data center.
RSD involves the establishment of various types of nodes, such as compute nodes, storage nodes, memory nodes, etc. These nodes are independently racked and contribute their resources to a pool. RSD dynamically assigns these resources to workloads and returns them to the pool when they are no longer required.
It is important to note a major disruptor of 2020 thus far, and that is COVID-19. The pandemic has impacted just about everything in our lives. Given this, technology is playing a big part in how our world continues on a day-to-day basis. More people are working from home and shopping from home, and this is all driven by technology. It is only fitting that all of this uncertainty and upheaval would have some affect on IT trends as well.
According to an IDC article in early May, overall IT spending around the globe will decrease by 5.1 percent due to the virus. Additionally, IT infrastructure spending is expected to increase by 4 percent. This means that whether organizations are moving their data centers from on-premises to cloud-based, vice-versa or creating hybrid environments, the need for IT infrastructure is crucial in 2020.
WWT is here to help you navigate the best approach for your organization's IT infrastructure. Our subject matter experts are available to answer your specific questions. Contact your WWT account team today to discuss in greater detail.
Alternatively, feel free to explore more data center and server infrastructure content for detailed information regarding the major original equipment manufacturers (OEMs) to consider.