Editor’s note: As an industry group that includes Red Hat, Cisco and IBM formed on Monday to push Software Defined Networking, more attention is becoming focused on this emerging technology. In fact, SDN’s increasing promise has been identified as a major reason for Oracle’s announced acquisition of Morrisville-based Tekelec.
To better understand SDN and what it means for the networking industry, WRALTechWire asked Mark Johnson, chief technology officer at MCNC and a leader in such projects as Internet 2, to offer his insight and analysis on this hot topic.
RESEARCH TRIANGLE PARK, N.C. - Software Defined Networking (SDN) is the latest buzz phrase in network technology. Recent acquisitions in this space signal what is likely to be a trend in the evolution of data centers as well as “cloud” technology.
The central concept of SDN is decoupling the moving of data from the decision about where data should move (network control). By separating these functions, software can be written that “virtualizes” the network, enabling centralized decisions to be made in software on the fly to move data in response to any external requirements. This enables enterprises to control movement of data within their own networks in more sophisticated ways and is likely to be a valuable innovation in data center environments.
Oracle recently announced that it is acquiring two companies actively developing SDN products; Tekelek, a maker of data center signaling products, and Acme Packet, which has a service delivery (especially unified messaging) product portfolio.
Oracle is positioning to be the underlying technology provider for cloud services both in the enterprise and for commercial service providers. It acquired SUN Microsystems in 2010 to add computer hardware to its core business of database software (and eliminate a competitor). With the acquisition of Tekelec and Acme Packet, Oracle is moving closer to owning all of the component pieces necessary to offer sophisticated data center or cloud applications services.
Oracle’s expanded product set is moving back to a mainframe world in some sense.
Large businesses once ran their software on huge mainframe computers with system software that controlled the movement, storage and selection of data. In the 1980s, the industry moved to a decentralized approach where many smaller computers would handle individual applications, movement of data (switches and routers), storage, and retrieval of data. Now we are seeing large enterprises and service providers that want to gain economies of scale and have more control over their own environments. SDN is a technology that enables this movement.
Oracle has positioned itself to be the one-stop shopping center for all of the components and will be able to offer tightly-integrated software solutions to manage everything in the data center.
Oracle is the largest and has the most complete product set of the companies in the marketplace, but they will not be without competition. Because SDN is a concept and not a product, companies that want to be innovative are likely to buy components that support SDN programming interfaces and build their own controlling software optimized for their specific needs.
While Oracle’s acquisitions are an aggressive move to position the company as a leader in this space Cisco, the largest maker of traditional networking gear, may have the most to lose as SDN takes hold. Cisco’s secret sauce has been their software. SDN opens up the software of networking to the end user and could completely disrupt the balance of power in the industry. Cisco has been setting up startup environments to stimulate innovation in this area. The sheer size of Cisco works against it as it plays catch-up with smaller, more agile competitors.
MCNC is looking to connect to a national SDN-enabled network operated by Internet2 and use that to both support our local network research community and to explore how SDN can enhance the service we provide to other types of research and the business needs of the broader set of MCNC users. We want to understand how network programmability can be made available to the users of the network and how it interacts with ordinary Internet use. SDN offers the promise of enabling users to define network capability and configuration on-demand which could improve performance and reduce cost at the same time. This is a compelling motivation for us.
Locally, researchers like Jeff Chase, a computer science professor at Duke specializing in utility or cloud computing, and Ilia Baldine, who leads RENCI’s network research and infrastructure programs, are at the forefront of developing SDN technology and exploring its application. MCNC is providing infrastructure in support of their research projects and the State will be well-positioned to take advantage of this technology as it moves from research to reality.
