CIO Virtual Roundtable: The Role of SDN in Hybrid Multi-Cloud Data Management

Most large organizations have come to the conclusion that they will continue to operate in an environment in which data centers with legacy systems coexist with private cloud and multiple public cloud resources.
As leaders explore ways to automate the management of this hybrid environment, many are looking at SDN as a way to virtualize and integrate disparate assets in a scalable and manageable manner.
SDN concepts are tying together today’s increasingly distributed enterprise infrastructure.

Most large organizations that have been around for a while have come to the conclusion that they will continue to operate in an environment in which data centers with legacy systems coexist with private cloud and multiple public cloud resources. As leaders explore ways to automate the management of this hybrid environment, many are looking at SDN as a way to virtualize and integrate disparate assets in a scalable and manageable manner.

Over the course of a recent CIO.com virtual roundtable co-hosted with Nikos Anerousis and Sean Burke of IBM, a dozen technology executives discussed the evolving role of SDN in advancing technology modernization initiatives across a range of industries.

Here is what they had to say:

An immediate consensus was reached around the room about how SDN concepts are tying together today’s increasingly distributed enterprise infrastructure. The rapid emergence of hybrid cloud computing, it was noted, has elevated the role of enterprise networking to provide a common fabric for managing workloads spread out across heterogeneous environments. As one participant observed: “Software defined networking definitely is part of our strategy as we consume more resources in the cloud while maintaining appropriate communication with our on-premises environment.”
It is an important line of communication as attitudes about third platform-based computing mature. Several participants remarked on the diversity of performance requirements across critical applications. In evaluating what goes to the public cloud and what stays on-premises, IT staff is wrestling with how to automate the process of putting the right workloads in the right place while tying them all together to enable effective data driven decision-making. This is especially critical when managing latency-sensitive applications and operations. “That is why edge computing has emerged as an imperative. When combined with SDN, it makes it possible to keep processing capabilities close to distributed data sources,” explained one executive.
Manufacturing, noted IBM’s Nikos, is a good case in point. “The way we see edge computing, is that it is about processing data by enabling computation at the IP edge. For example, on factory floors or in warehouses, a lot of data is being generated by devices and machines locally. It can be super expensive to send all of that data to the cloud.”

Moreover, sending a high volume of this type of data to the cloud and back is too slow. “You cannot do real-time robotics or visual quality inspections, etc., using traditional manual methods. It requires a very different computation architecture that supports the installation of processing capacity near where it is needed.” SDN introduces the intelligence and automation needed to harvest value from the edge while integrating and sharing key insights with the right people and applications across the enterprise.
Traditional networking paradigms are also going to be difficult to maintain as organizations embrace the concept of “infrastructure as code.” As organizations rapidly deploy applications across public and private cloud resources, several participants noted the need to integrate DevSecOps principles and models into their network management operations. “If you are deploying and refreshing applications across the enterprise several times a day, it will be difficult to do this using traditional structures.”
While discussions about hybrid infrastructures have tended to focus on the complexity associated with managing on-premises and cloud resources, several executives noted that there is plenty of heterogeneity within enterprise networks themselves. Organizations are integrating Zigbee, Bluetooth, local area networks and wide area networks while continuing to support existing investments in legacy networks.

As one executive observed: “We are in the process of figuring out how much of our current MPLS network environment we can actually take down from our various locations. We are analyzing what makes sense.” So, just as legacy and modern computing resources will coexist for the foreseeable future, legacy and modern networks will have to figure out how to live together in a constructive manner. SDN is emerging as a “quarterback technology” for orchestrating these hybrid networking environments.
Integrating analytics and automation across hybrid multi-cloud infrastructure elements (including legacy and modern networks) will require SDN capabilities. According to IBM’s Sean, there are mission-critical IT strategies that are hastening SDN adoption. For instance, as organizations assess the role of edge computing -- and the ability to work with highly distributed data -- leaders will have to consider how enterprise networks can deliver compute capacity closer to the source. “Where are you performing the analytics in this complex situation? SDN provides a tool for managing this Frankenstein-like environment.”

To read more, or for more information on participation, please visit www.CIO.com.