Usage based billing, bandwidth throttling, a pay per use internet. Sandvine’s business can described with a host of hot button terms. Or it could be described with just one: Netflix. The growth of Netflix is bringing debate about bandwidth management to a head. On May 17th, Sandvine released a report showing that Netflix now accounts for 29.7% of all downstream traffic during peak period. Waterloo’s Sandvine sells technology that gives service providers a window into the world of their chaotic traffic. Many investors believe Sandvine may be able to help US service providers avoid the worst scenarios, or at least curtail the problem until more capacity becomes available. Recently, Cantech Letter talked to Sandvine President and CEO Dave Caputo.
Dave, can you give our readers a bit of an overview of Sandvine’s history?
Sandvine was established in 2001 by the same management team that built PixStream and sold it to Cisco for a half-billion dollars. Like Sandvine, PixStream was a networking equipment company. Sandvine’s original success came through the development of a Network Policy Control solution that managed the upstream congestion problem faced by cable operators of the day. Gnutella and other music and file sharing programs were transforming the Internet from a tool where end users only downloaded content to a bi-directional network where end users also uploaded data. Cable networks were designed to receive TV signals downstream so the bandwidth available to accommodate files uploads was minimal. Sandvine intelligently managed the traffic to make better use of available bandwidth. Today, we still have most of North America’s largest cable providers as customers. Since then we have significantly diversified the business. Telcos that offer DSL services are now the largest component of our customer base and we have very quickly grown a substantial base of mobile service provider customers, a group that wouldn’t talk to us a couple years ago. Today we have well over 200 Internet service provider customers in more than 80 countries. Over half of our business now comes from markets outside of North America. We have also have a much more broad solutions portfolio today. While we still manage network congestion, we also have solutions that help service providers devise, create and enforce new service plans for subscribers. We can measure network quality, mitigate malicious traffic, like Spam, and enhance the efficiency of third party solutions, like network caches, by directing only the relevant network traffic to them (e.g., only cacheable traffic) – all from the same intelligent hardware platform. Our revenue last year was approximately USD$90 million – 48% higher than USD revenue the year before.
What is deep packet inspection?
Deep Packet Inspection, or DPI, is just a passive inspection technology. It can identify characteristics about Internet data packets, such as its protocol, like http (Web browsing), or RTSP, a media streaming protocol. The real question is what can DPI do? DPI is a major supporting technology for Network Policy Control, which is a class of solutions that helps Internet service providers intelligently manage and control their networks for an optimal subscriber experience and increased efficiency and profitability. For example, When the network is congested, the Internet is designed to drop or slow packets to make room. But the Internet does not make these decisions intelligently. Data from the voice call you are on at that moment are just as likely to be affected as data from an unattended file download, say for a Microsoft software update. You care whether you can understand your phone call. Do you care if your software patch takes a few moments longer? A Network Policy Control solution could automatically make the right decision every time. Or perhaps you have signed up for the all-you-can-eat social networking tier from your mobile service provider. You can send and receive updates to and from Facebook all day and night for one fixed price. In Latin America, where Facebook alone represents 13% of traffic during peak network hours, a Sandvine mobile network customer has created just such a service plan that has proven to be a real competitive differentiator in their market. A Network Policy Control solution sits behind that plan to identify the social networking traffic and to communicate with the billing system to “zero-rate” all such traffic.
On May 17th, Sandvine released a report showing that Netflix now accounts for 29.70% of all downstream traffic during peak period. This is up almost 10% from just last October. In the report you say that “the age of Internet video is upon us”. What does this mean, practically, for consumers and for Sandvine?
That’s correct. In fact, even when measured over a 24-hour period and looking at ALL traffic – downstream and upstream – Netflix is the single largest component of traffic in North America. In the past seven months, Netflix has added 6.7 million subscribers and has 23.6 million subscribers, as of their last quarterly report. In Canada, in the four months that followed its launch, Netflix grew to 14% of downstream traffic during the peak evening period. There were 900,000 subscribers in Canada according to Netflix’s last quarterly report. The success of Netflix’ expansion into Canada suggests that it is not a phenomenon localized to the United States. With the company poised for continued international expansion, fixed and mobile Internet providers worldwide must plan for a future in which on-demand video (whether provided by Netflix or another service) is a large proportion, if not the majority of, last-mile traffic. Streaming video traffic can be sensitive to any latency – delays in delivering packets – in the network. Consequently, service providers need to build more intelligent networks that can distinguish between latency-sensitive traffic like certain video streams and bulk traffic like file sharing or email that is less time sensitive. Any latency sensitive traffic would need to be delivered with some priority when the network is congested or near-congested. Clearly the growth of online video is a major threat to the traditional broadcast video business of network operators, particularly cable. Sandvine offers these service providers executive-level dashboards to monitor the day to day impact of “over-the-top” video on their networks and their business. Service providers may want to differentiate themselves competitively by creating new service tiers that are targeted at heavy video users, so that their network connection characteristics are optimized for the video experience. Or, service providers may want to get into the Netflix business themselves. As long as the networks can deliver the kind of video experience that they are used to from broadcast TV, I think the trend is great for consumers. Why should anyone tell you when to watch your favourite TV programs or movies? We are moving to a real-time, on-demand world where the consumer controls his entertainment to a much greater extent.
Are bandwidth issues always going to be with us, or do you envision a time when we will have all the bandwidth we need?
To date, fixed line broadband has been a Field of Dreams – “build it and they will come”. No matter how much capacity is built, it is almost instantly consumed. The fastest networks in the world, in places like Korea and Japan, still face congestion. Instead of downloading a couple of movies a day, they download several. Or instead of streaming video in standard resolutions, they stream in HD. Microsoft just bought Skype, which is already the biggest long distance phone company in the world. What happens when Skype goes HD and long distance HD personal video conferencing becomes the norm globally? In mobile networks, bandwidth is limited by the availability of radio spectrum, which is itself a limited resource, so there will never be “all the bandwidth we need”. As a result, there will likely always be a great need for Network Policy Control in mobile networks. Even if we had a limitless amount of bandwidth, Network Policy Control would still be very necessary. For example, service providers will always want to differentiate themselves competitively through innovative service plans and those plans need intelligent solutions in the network. It will always be beneficial to mitigate the amount of malicious traffic on the network and Network Policy Control solutions can identify and reduce that type of traffic.
Sandvine’s streak of six consecutive quarters of revenue increases ended with Q1, 2011’s numbers. Can you resume the upward trend?
Revenue was below expectation in the first quarter as certain opportunities that we expected to materialize within those 12 weeks were delayed. However, many things went well in the first quarter too. Sandvine won a total of 13 new service provider customers — the highest level in over two years. Two of those customers were Tier 1 service providers. One of the customers was our inaugural win with our new reseller partner Acision, a major systems integrator in the mobile market. We did not lose any notable deals in the quarter. We responded to more RFPs, RFIs and RFQs in Q1 than we did in the fourth quarter of 2010 and the sales funnel has never been larger. We also announced our new Policy Traffic Switch 22000 platform, which is ideal for both fixed line and new mobile LTE networks. The PTS 22000 has up to 20 Gbps of performance in a highly cost-effective two-rack-unit form factor. Deployed at the edge, the new platform will be able to intersect certain network links that let us gather critical information to enable unique policy opportunities for service providers to run their networks more efficiently and increase revenue opportunities. We already have a first customer for this exciting new platform.
A real bright spot for Sandvine is sales to the Asia-Pacific region. Between 2009 and 2010 sales to the region as percentage of your overall revenue increased from 18.3% to 42.7%. Did you choose to focus on this part of the world, or is it choosing you?
Both. We have dedicated a great amount of time and resources to the area. Not too long ago we moved one of our top senior management team members to the area to oversee operations. We also have struck some key relationships with local partners, such as Mitsubishi, which are key to success in Asia. For success in Asia, you need local representation and a deep understanding of the business culture there.
As a result of these efforts we have won some flagship accounts in the region, such as NTT Communications, which is admired globally for the quality of their networks and operations. Accounts like NTT are tremendous references for other opportunities in the region and globally. We definitely expect to grow business in Asia over time as there are many more opportunities available there. On a quarterly or annual basis, it’s more difficult to predict what will happen as one or two large deployments in one part of the world can skew our results over such a short period.