Cisco Subnet An independent Cisco community View more

Interop and SDN: Overall disappointment, but some points of light

A follow-up explaining how the 10 important SDN trends I hoped to see at Interop actually played out.

For anyone who read my pre-show blog (Top Ten Advancements I Want to See at Interop Las Vegas) outlining my hopes (and, admitted, dreams) for SDN at Interop, you know I set the bar very high for a technology that represents a dramatic change from the status quo in networking. After all, in networking, evolution trumps revolution every time. Quite frankly, I actually expected to be disappointed. And I was.

That doesn't mean I didn't see positive movement forward for SDN. I did. It just came in the form of small steps versus giant leaps. Here's my top ten list:

10. HP-Microsoft Lync Server Proof of Concept. In my pre-show blog, I stated how I wanted to see SDN movement beyond the data center. This demo showed SDN at work across any portion of the network and highlighted SDN's potential for improved automation and network dynamics. The demo was simple - one to which we can all relate. It involved screen sharing. You know the drill. Share your screen with your remote partner(s) and they see what you see - only usually with a high degree of pixelation. Think how your cable TV looks when the set-top box or the service provider isn't quite doing its job. Here, an HP SDN application works with Lync Server to dynamically reconfigure the underlying network so the Lync Server session gains the necessary bandwidth to deliver a high-quality screen-sharing session. The remote partner(s) actually does see what the originating partner sees. Could you deliver this service quality without SDN? Yes. But it would take more knowledge, more work, and more time. Here, SDN supplies the knowledge, does the work, and saves the time.

9. NEC Application Center. I was able to spend some time at the SDN Workshop held at the show. My favorite moment from the workshop involved an audience member who asked, "When can we expect an SDN 'App Store' to be available?" The audience laughed... nervously. Why nervously? Well, on the one hand, everyone in the audience knew that such a single storefront is a networker's pipe dream. On the other hand, everyone in the audience also knows how great and wonderful such a single storefront would be for operators and the SDN movement overall. Talk about your SDN giant leap! Back to the small steps...NEC launched its SDN Application Center at the show. While NEC's Application Center will take time to populate, the fact that a vendor has stepped forward to enable the development, validation, and delivery of SDN applications that leverage its controller is a strong positive. Early SDN Application Center contributors include vArmour, Silver Peak, A10, Real Status, and Red Hat - all offering unique network services that complement an SDN infrastructure. Would we like to see a consortium of vendors leverage one single store with one set of development tools, certification criteria, and delivery process? Absolutely. But, for the time being, I'll settle for one vendor taking the lead.

8. Tail-f SDN Management Components. Tail-f has been supplying networking management software components to equipment suppliers for years now. This saves the supplier time and resources, and results in a more functional and solid solution for the supplier's customers. What does this have to do with SDN? Well, two core tenets of SDN are network simplification and automation. Tail-f's two products - ConfD and Network Control System (NCS) - go right at these core tenets. ConfD device management software is provided to networking suppliers looking to develop systems that meet the increasingly complex and ever-changing requirements relating to configuring, operating, and adapting network devices. NCS is provided to network operators looking to consolidate and automate the management of both network devices and services. Anyone who was at Interop saw NCS at work in the InteropNet. That's a tough test for any management system.

7. Plumgrid Virtual Network Infrastructure. The link between SDN and Cloud Services is strong. The elastic nature of the cloud is enhanced by the flexibility of an underlying SDN infrastructure. And nothing is more flexible than a virtualized environment - server, storage, or network. Plumgrid takes virtual networking to a whole other level by offering a complete network (switching, routing, load balancing, firewall) in software. VNI offers the cloud services subscriber a segmented singular network - with services that can be adapted to individual subscribers. And, as would be expected of a cloud-centric virtual SDN solution, VNI can be controlled by a selection of hypervisors and orchestrated via an OpenStack Quantum plug-in.

6. Jeda Networks Software-Defined Storage Networks. In terms of service and operational advancements, server virtualization has outpaced storage virtualization. And both have outpaced network virtualization. Many current and planned SDN solutions advance network virtualization. Will these advancements leave storage lagging behind? Not if Jeda is representative of things to come for storage virtualization. Leveraging many of the SDN technologies and techniques being applied to the network, Jeda is developing storage solutions that truly separate storage logic from storage devices. The result: Increased agility and efficiency - two primary benefits of SDN in the network. Looks like as LANs and WANs go with SDN, so too will SANs.

5. SDN Hardware. While some vendors (i.e., software vendors) are promoting SDN as THE technology that renders all hardware developments in networking moot, there are many other vendors looking to optimize hardware platforms in preparation for the SDN future. While SDN advancements will undoubtedly move many networking functions (e.g., control systems and network services) to software running on standardized servers and appliances, my view is that hardware platforms designed and dedicated to networking will continue to flourish in the future. At Interop, many hardware component manufacturers (e.g., Broadcom, Intel, LSI, and Netronome) were showing off their smaller, faster, and smarter chips aimed at handling the increased decision-making and improved dynamics that will characterize the SDN environment.

4. SDN Testing. SDN offers much potential for networks and network operators. Witness supplier developments, operator interests, standards efforts, lead user deployments, and financial commitments. What SDN lacks right now is proof that it can deliver on all its promises. Comprehensive testing of multivendor SDN solutions is one of the primary proving grounds for available and developing products. On the testing front, two vendors stood out for me at Interop - Ixia and Spirent. Ixia was the first test system vendor to join the Open Networking Foundation - the driving force behind OpenFlow advancements. Ixia also plays a lead role in the ONF PlugFests. The 2013 PlugFest is slated for June 3-7 in Indianapolis. Here, interoperability tests of OpenFlow 1.0, 1.2, and 1.3 are planned. Given that ONF has stabilized on OpenFlow 1.3, pay close attention to these test results - and participating suppliers. (Oh, if the test engineers could talk?) While not specifically aimed at SDN, Spirent's release of an appliance-based testing product puts easy-to-use "what if" testing in the hands of network operators. With so many operators looking to ease into SDN and having limited visibility into network traffic, tools that exercise the network before actual deployment of new solutions, services, and/or applications reduce risk and improve outcomes.

3. Juniper's Contrail SDN Controller. I'll admit to being rather critical of Juniper's SDN coming out party this past winter. In the announcement, I saw lots of big boxes, nice linkages, and service chains, but little in terms of specifics. As you see by my previous NWW SDN blog, I think SDN has reached a point in its development where more substance and less style is needed. Enter Juniper's Contrail solution (at least by end of year). Acquired by Juniper late last year, Contrail functions as central SDN command and control. How does it differ from the many other SDN controllers out there? Well, for one, it brings integrated analytics capabilities to the mix. Without solid network state inputs, the accuracy of network reprogramming could be called into question. Second, it closely couples network virtualization and network orchestration. Abstracting the network reduces complexity. Abstracting AND adapting the network together not only reduces complexity, but also drives improved network service quality and resource utilization.

2. FIDO. Admit it. How many of you don't know what FIDO is? FIDO represents the future of user authentication. Think automation. Think single sign-on. Think of all those login tasks that drive users crazy and keep administrators nervous. While not close to being labeled an SDN security service, FIDO's automated approach should certainly serve as a nice complement to a simplified and automated SDN environment in the future. Think about it... Every login takes place across a network connection. It would seem to make sense to link FIDO and SDN at some point.

1. The New Networker. Most of you reading this blog are likely dedicated networking professionals. Many of you have many valuable network certifications on your resume. While I place a high value on all this network expertise and experience, I would assert that SDN will increasingly demand networkers to expand their expertise into the computing side of IT. SDN is already drawing computers into the network. In talking with Interop attendees, it was quite evident that those most excited by SDN were those looking to improve their whole IT infrastructure - versus those looking to improve only the network itself. This "systems" view is likely on display in your own organizations. You may have noticed more questions about SDN coming from the computing side of your organization. You may have noticed more demands for you to learn computing components - e.g., OpenStack, KVM, APIs. Is the network manager a dying breed? Not at all. Network knowledge is absolutely critical to "systems" success. Let's say the network manager is evolving - just as all other IT managers are.

Was I disappointed there were no giant leaps in SDN at Interop? Yes. But I was also encouraged to see the above small steps in the right direction. What's that they say about any long journey beginning with a single step? With SDN, we seem to be seeing more of these steps taken every week... moving us ever closer to our end goal of a simpler, more efficient, more dynamic, and, most importantly, more valuable network infrastructure.

To comment on this article and other Network World content, visit our Facebook page or our Twitter stream.
Must read: Hidden Cause of Slow Internet and how to fix it
Notice to our Readers
We're now using social media to take your comments and feedback. Learn more about this here.