Making Cloud Architectures Work
For the enterprise, virtualization and the cloud have become reality, and software defined data centers aren't far behind. But for all their touted benefits, these architectures also create plenty of challenges. I spoke with Mariano Maluf, president of the VMware User Group (VMUG) and cloud ecosystem architecture lead for Coca-Cola, about what organizations need to consider to make virtualization, cloud architectures, and software defined data centers work.
Optimizing virtualized deployments
The virtualization of the data center is proceeding apace, Maluf said, adding that "a lot of companies large and small have achieved 70 to 80 percent virtualization in their data center." Virtualization itself is only part of the battle, however. Getting to the next level, where that virtualization provides the maximum benefit, requires enterprises to answer several pressing questions.
"How do we simplify the lifecycle management of virtual machines? How do we get smarter at analyzing logs and doing triage on problems? How do we correlate data through the different layers of the virtualized data center so we can troubleshoot faster and fix problems sooner?" Maluf asked. In many cases, the answer to those questions won't involve more management, but rather more automation.
"Policy-based automation at the heart of the software defined data center is the right way to go," Maluf told me.
That automation, moreover, needs to pervade "the whole infrastructure stack, not just the compute layer," he said. This is what will optimize virtualization and SDDC deployments and provide the business agility the next generation of architectures promises.
Convergence and the future of the networking pro
To get there, organizations will have to change the way they think about IT. They must now consider how to bring the disparate elements of their architectures together. The silos of legacy IT won't work anymore.
"The silos of the past were very successful at providing certain capabilities for certain business units or landscapes, but they progressively became a group of islands. Private cloud infrastructure requires a different mindset from an organizational perspective, from a people skills perspective, and from a process perspective," Maluf said. Mindsets must shift away from nuts and bolts, towards how data center operations and IT as a whole can contribute to business goals that benefit the enterprise. Convergence, in other words.
For networking professionals, that convergence entails a profound shift in thinking, according to Maluf. "Think back to when networking was really mostly physical. You needed to connect cables and build the connectors for specific ports on servers, and so on. Then we started to change the paradigm. Virtual networks appeared, and there was a shift that network professionals had to do. What we're seeing now is deeper," he said.
In response to this shift, network professionals should begin to consider incorporating more advanced automation and virtualization principles to their skill sets. Some already are.
"What we see in our community at VMUG is an increased interest from network professionals in attending our sessions and getting familiarized with this messaging," Maluf said. And with the launch of VMware's NSX network virtualization solution, which he said "has the potential to be a true game-changer," networking pros' interest in VMware and virtualization should only continue to grow.
Software defined data centers and network virtualization look set to be hot topics for some time to come. "When companies are looking at their highly virtualized data centers and seeing what's next, the opportunity is there to take this software defined data center architecture seriously," Maluf told me. The question, he said, is "How do we really make this cloud architecture work?" With smart planning, networking professionals can make sure they are key to answering that question.
Jude Chao is executive editor of Enterprise Networking Planet. Follow her on Twitter @judechao.