Like a traditional data center, a software-defined data center (SDDC) houses servers, storage, networking devices and other IT or telecommunications hardware. However, in an SDDC all those infrastructure elements are virtualized and delivered as a service. In other words, the software is abstracted away from the hardware, which allows organizations to manage their compute, storage and networking resources as virtual pools. Instead of having to configure hardware manually, staff can control their infrastructure from one central software interface.
This type of architecture not only saves IT staff time, it also gives organizations more flexibility. In addition, it makes it easier to use inexpensive, commodity hardware rather than integrated solutions that can lead to vendor lock-in.
While the concept of virtualization has been around since the 1960s, the term "software defined data center" first came into common use in 2012. Steve Herrod, then the CTO at VMware began using the phrase to describe the evolution of the data center into an environment where everything was virtualized.
In the years after its introduction, the SDDC concept experienced a dramatic rise in popularity. By 2015, market research firm Gartner declared "the future of the data center is software-defined." In the same press release, the firm predicted that by 2020 "the programmatic capabilities of an SDDC will be considered a requirement for 75 percent of Global 2000 enterprises that seek to implement a DevOps approach and a hybrid cloud model."
Elements of a Software-Defined Data Center
There is no single piece of hardware or software that an organization can deploy in order to achieve a software-defined data center. Rather, SDDC architecture is a complex approach to data center management that requires the integration of many individual solutions.
While each software-defined data center will be slightly different, in general, SDDC architecture includes three basic layers: the physical layer, the virtual layer and the management layer. The physical layer is the actual hardware — the servers, storage and networking gear that takes up space in the data center.
The virtual layer is the software that abstracts each of these resources and delivers it as a service. It includes the hypervisor that manages the virtual machines (VMs) or, as is becoming more common, the container software. In addition, it includes the software-defined storage (SDS) and software-defined networking (SDN) software. It may also include other virtualized services, such as virtualized security.
On top of that is the management layer that ties everything together. This may include a software-defined data center solution, orchestration tools and/or automation capabilities that make it possible to run the data center from a centralized interface.
A software defined data center consists of three interlocking layers.
Managing a data center as an SDDC offers a number of benefits compared to traditional data center management:
- Time Savings— Because SDDC managers can configure, monitor and maintain their systems from a centralized interface, they no longer have to walk around and physically touch the actual hardware in order to accomplish most of their day-to-day tasks. Automation handles many jobs that would otherwise need to be done manually. As a result, managers spend much less time on these activities, freeing them up for other responsibilities.
- Cost Savings — SDDC architecture can save companies money both in terms of capital expenditures and operational expenditures. On the CAPEX side, software-defined infrastructure can run on commodity hardware, which in general is less expensive than proprietary systems. In addition, resources like compute, storage and networking are pooled together. That allows organizations to use these resources more efficiently rather than overprovisioning to account for peak demand periods. In terms of OPEX, as already mentioned, the automation in an SDDC reduces the amount of time that staff spend maintaining the systems, which helps reduce total cost of ownership (TCO). In addition, because many organizations deploy converged or hyperconverged infrastructure in their SDDCs, they also frequently save on real estate and energy costs as well. Some studies have found return on investment (ROI) for migrating to an SDDC can be as high as 220 percent.
- Avoid Vendor Lock-In — For many organizations, the biggest appeal of SDDC is that they can use infrastructure from any vendor rather than getting locked in to a single supplier. This can result in additional cost savings as well as allowing companies to take advantage of technological advances as they come into the market.
- Flexibility and Agility — Markets are constantly changing, and businesses shift in response in order to take advantage of new opportunities. Often this requires corresponding changes in IT infrastructure, and SDDC architecture allows organizations to make these alterations much more quickly than would traditionally be the case. Instead of needing weeks or months to provision servers and storage for a new application, IT with an SDDC can often perform the task in just minutes.
- Reliability — Because resources are pooled and workloads can be shifted as necessary, an SDDC is also more resilient than a traditional data center. If a particular piece of hardware fails, the management software can simply move the workload to a functioning piece of infrastructure — often without the need for manual intervention. In addition, the most advanced SDDC management software may include analytics capabilities that can alert staff in advance when problems are likely to occur so that they can take action proactively.
- Analytics and Insights — In an SDDC, monitoring tools are constantly collecting log data about the performance of various systems. That opens up the possibility of using modern analytics tools to comb through the data for insights that can help optimize the systems for more efficient and cost-effective operation.
Of course, migrating to an SDDC also carries risks and potential problems that must be overcome, including the following:
- Complexity— Many organizations say that integrating all the disparate solutions that comprise an SDDC is one of the biggest headaches involved. For most firms, migrating is a slow, complicated process and getting all the software and hardware to work together as intended can take a long time. As a result, many firms choose to contract with a consultant or integrator in order to speed up the process and ensure that things go as smoothly as possible.
- Security — In a survey conducted by vendor HyTrust, security was IT leaders' number one concern related to the cloud and SDDCs. The complex nature of an SDDC can make them more difficult to secure than traditional infrastructure. However, in the same survey, a majority of respondents felt that the SDDC security tools currently on the market were up to the task.
- Legacy Hardware — Most organizations do not want to "rip and replace" their old infrastructure when migrating to an SDDC. After all, part of the point of virtualized infrastructure is that it can run on nearly any hardware. In practice, however, organizations often find that newer hardware, such as converged or hyperconverged systems designed for SDDCs, is easier to incorporate into the software-defined environment. While it is possible to integrate older systems, some organizations find that the time involved makes it impractical.
- Hybrid and Multi-Cloud Environments — Many organizations with SDDCs also have private clouds and use multiple public cloud services. Setting up a hybrid cloud adds yet another layer of complexity that further complicates the tasks of integrating and managing the various systems.
SDDC and Cloud
While it is possible to have an SDDC that isn't part of a cloud, and vice versa, the two often go in hand. SDDC and cloud computing have a lot of overlap. Both deliver infrastructure as a service, and both rely on pooled and virtualized resources. In addition, sometimes the same management and orchestration tools can run both a cloud and an SDDC.
Organizations often deploy private cloud and SDDC technology at the same time because the two are so complementary. In other instances, organizations will first begin rolling out SDDC capabilities as a stepping stone to the private cloud, while firms that already have private clouds sometime embrace SDDC architecture as a way to improve manageability and further reduce costs.
Future of SDDC
In general, organizations that begin adopting SDDC architecture do not roll it out to their entire data centers at once. Instead, they usually deploy it in stages, often as part of their regular hardware refresh cycles.
As organizations experimenting with SDDC begin to see more benefits from the approach, analysts believe that it is likely that SDDC technologies will experience rapid growth. Trends like cloud computing and DevOps could also accelerate adoption. Allied Market Research says that the SDDC market is growing at an average 32 percent per year and predicts that the market will be worth $139 billion by 2022.
Two other technologies are likely to affect SDDCs: containers and artificial intelligence (AI). As Docker and other container technologies become more popular, it seems likely that future SDDCs will have more containers than VMs. And AI could augment the analytics and automation capabilities of SDDC management tools, further reducing the need for manual human intervention.
In short, the future of SDDC looks likely to hold even more growth and technological innovation that will enhance the benefits to organizations.