Scanrail - Fotolia
Naturally, some companies are hesitant to adopt a software-defined data center, as overhauling existing platforms in order to use a single vendor's suite can be extremely costly. If the company isn't in a position to move to a single vendor, another option is to keep some legacy platforms and replace others. There are drawbacks to this option, too. By using platforms from different vendors, you risk some level of incompatibility. Also, it can be challenging in terms of obtaining the necessary skill sets to effectively oversee these environments.
As always, the first step to making a big decision like this is to do your research. We caught up with the EMEA solutions specialist for VMware at Hitachi Data Systems, Valentin Hamburger, who gave us some expert insight on software-defined data centers.
What are your best practices for building and managing a virtualized architecture?
Valentin Hamburger: Start with a clear scope and note all requirements, limits and constraints. Develop a virtualized architecture that is satisfying the scope -- requirements -- and can work with the given limitations and constraints. It's worth revisiting the design once in a while, since requirements might change over time.
What skill sets are necessary to build and manage a virtualized architecture?
Hamburger: A profound understanding of IT processes and business requirements will definitely help. It's often necessary to translate business requests into IT projects. Therefore, a basic understanding of the business and its challenges is tremendously helpful when running a virtualized and automated data center.
Is there a difference between a software-defined data center and a private cloud?
Valentin HamburgerEMEA solutions specialist for VMware at Hitachi Data Systems
Hamburger: This highly depends on how the term private cloud is defined. If it's a data center owned by a company, the software-defined data center and the private cloud are quite the same. One could say that the software-defined data center is the foundation for the private cloud -- to deliver cloud-like features and functionalities. However, some people define the term private cloud as a part of a public cloud provider -- tenant -- except that only one company can exclusively access it. Then, the software-defined data center in the local data center might be configured to create a connection to that provider-hosted private cloud to form a hybrid cloud experience. This setup would be able to provide the best of both worlds -- data center locality where necessary and extensive cloud flexibility with a single software-defined data center management interface, otherwise known as cloud brokering.
How important is it to incorporate flexibility when building software-defined data centers?
Hamburger: The software-defined data center is about enabling the business to be more agile and to react faster to their daily market challenges. To do this, flexibility is a must. An inflexible software-defined data center might add some automation benefits, but it will end up not being used if it can't fulfill business demands. If there is a waiting time on resources for several months, the business will look for an external alternative to fulfill its needs quicker.
What are some of the challenges customers face in building software-defined data centers?
Hamburger: One of the biggest challenges is processes. These are nontechnical issues and they can only be overcome by including all parties in a software-defined data center project. It must be a corporate joint venture where all departments work together to make it successful. Changing processes is necessary for a software-defined data center to work -- this should be embraced, as it's a huge opportunity to enhance the old legacy ways of doing things and introduce the cloud/automation/agility approach to any enterprise.
What would you to say to companies hesitant to implement a virtualized architectures?
Hamburger: It depends on the use case and the scope. Typically, it's realistic to say that about 90% of workloads in a data center can be virtualized just fine. For the remaining 10%, the reason why they can't be virtualized might not be technical. There are licensing issues with some vendors that render virtualization so expensive that it's cheaper to run the software on physical servers. However, if licensing isn't the problem, but performance -- there are very rare circumstances where a physical environment outperforms a virtual one. In general, virtualization adds so much more than simple consolidation to a data center that it's hard to imagine a modern data center could work without it.
How do you think business needs and software-defined data centers will continue to change in the future?
Hamburger: I guess the business will get even more demanding in the future. We, as a society, are used to getting what we want quickly and easily as a service. However, the back end still needs to be client-server solutions, which now need to scale based on an unknown user behavior -- will 10 or 1000 people use the service today? Based on that development, the whole hybrid use case will become far more important -- since capacity can be quickly "rented" for as long as it's needed and then given back if it becomes unused again.
Further in the future, there might be giant data centers delivering IT infrastructure for customers. Those customers will define what they're going to run on top of this infrastructure on a pay-per-use model and how much resources they need. It might sound like data center housing, but it will be far more flexible and accessible. This will make it far easier for enterprises to create new projects or build new business cases -- since it can be tested and achieved without any hardware costs.
If you'd like a bit more information on making the switch to a virtualized architecture with a software-defined data center, read an excerpt from Building VMware Software-Defined Data Centers, by Valentin Hamburger, on why modern businesses need this technology.
Move toward a software-defined data center via autonomic computing
Incorporate integration and scalability in your SDDC plan
Test your knowledge of software-defined data center trends