Cloud/SaaS/SDN, Forum

ConteXtream Seeded by BGV – By Eric Benhamou

In 1994, a long time before Israel became known as Start-Up Nation, I met a young company named NiceCom, an early technology leader in the emerging field of ATM switching. The company was then a subsidiary of Nice and was lead by Nachman Shelef. Sharon Barkai was one of his most creative engineers. At the time, I was running 3Com. We decided to acquire NiceCom that year. This was the transaction that ignited interest in Israeli high tech and unleashed a floodgate of successful start-up exits in the field. Twelve years later, I hooked up with Nachman and Sharon in the lobby of the Tel Aviv Hilton. That time, we discussed their latest ideas about re-injecting innovation in the dormant, Cisco dominated, networking industry. The concepts they were describing to me had true break-through potential. I decided to invest in their budding company ConteXtream on the spot. This is how BGV became ConteXtream’s first seed investor. I talked to Sharon again this past week to reflect upon the path traveled since our initial meeting.

ConteXtream Seeded by BGV

A conversation with Sharon Barkai, founder and CTO

1- When you and Nachman Shelef first started ConteXtream, what did you think was the most critical problem you wanted to address? What was the breakthrough idea that got you started? One of the key challenges of ConteXtream was actually that the company was established based on observations regarding the possibilities of new networking architectures, and not any specific problem in good old IP / Ethernet bridging & routing. These observations were about tasking (distributed) software with the job of .. building a network, or connecting things. The immediate implications from this idea were: a) we need a basic IP infrastructure in place so we can distribute software-defined anything and b) a software-defined network mapping function has extraordinary identity based qualities and abilities that eliminate traditional networking complexities. It was only later on that the true applications for software-defined networking presented themselves. These had a lot to do with the separation of the sequential client hardware (e.g. mobile smartphone) from the (multi-core) concurrent server hardware, the formation of elastic clouds, the mobility of virtualized environments, and cloud networking. 2- How has your original vision evolved over the past 5 or 6 years? Which part of your original vision has remained the same, and which part has changed? The basic vision has remained mostly intact and, while it has begun to materialize, has not yet fulfilled its full potential. But we had to constantly evolve its communication, both externally and internally and connect the dots between the current state of the market and where we wanted to go. For example I recall one of our early marketing offsite meetings moderated by an iconic Silicon Valley communicator. It lead us to refocus our pitch to investors entirely on network virtualization. Needless to say reaction was not very enthusiastic … It was premature, virtual machines were not yet a well known concept, neither was vmotion, elasticity, or hosting, and carrier network function virtualization was definitely beyond the “after we retire” horizon. Naturally all that has evolved after more innovative players coined the terms SDN & NFV and started articulating related notions. 3- SDN started off as a way to open network innovation in research environments, but it has fast become a buzzword and a hot investment area. ConteXtream deserves credit for being a pioneering inventor of its fundamental concepts. Can you separate the actual reality from the fiction? Which types of large scale SDN deployments are possible today? What do you expect over the next couple of years? The emergence of the SDN discussion in the industry definitely made our lives simpler in terms of communicating the kind networking architectural innovation we wanted to show. Before that milestone, there was a sense that growing the IP transport infrastructure had to be the end-all play, given the success of broadband and the Internet. Yet, as compared with the early SDN “marketechture” descriptions, we still looked a bit different. The original SDN proposals portrayed an out-of-band network software controller as if it was a smart robot controlling a Detroit Auto-Assembly line. It is only in the past 12 months or so that we have witnessed a growing realization that SDN needs to extend — not replace — the IP transport infrastructure, and that federated emergent architectures such as ConteXtream’s distributed flow-mapping design are the way to go about it. 4- Recently, there has been growing support among carriers for Network Function Virtualization (NFV), notably within ETSI. What are your views of the principal goals and benefits of NFV? How does NFV’s focus differ from or complement SDN? To us the NFV trend completes the picture of what we started day one. It is not only about the ability to network and connect identities using distributed software and distributed flow-mapping, but also about making every function carriers want to apply to their traffic from the access to the core and back be instantiated the same was as a Google search, in other words enabling these giant communication service providers to operate just like Internet companies! I am referring to their capex efficiency, their service velocity, their time to market — even though they don’t write their own apps , and they operate using a much more geo-distributed set of points of presence. 5- Why should mobile carriers embrace SDN today? Mobile carriers connect the things that matter most to consumers, their personal device, phone, tablet, car, watch.. to the cloud. This is the most competitive, dynamic, and demanding networking environment today. Innovation is needed in every aspect of this business to ensure survivability and success. If there is a service optimization or monetization opportunity that your competitor can “weave” into the network more quickly or more efficiently than you, then you may not be able to live to play another round. In essence this is the capability that SDN or network function virtualization brings to the table. It functions in the direct line of traffic and complements the virtual compute-storage environment which was introduced in the past decade. It is a win-win-win for the carriers, end-customers, and technology innovation vendors. ConteXream already today is able to offer its customers in production the ability to “chain in” a menu of functions, and to efficiently load balance and service match a 50 million white-pages subscriber base to the yellow-pages set of functions, 50 billion times a day! I think we could say this is definite evidence of SDN has reached prime time with mobile carriers.
Sharon Barkai, ConteXtream Founder and CTO

Sharon Barkai, ConteXtream Founder and CTO