Network analytics startup Saisei, of which Dodo founder Larry Kestelman is the largest shareholder, has been recognised as a ‘Big50 Startup’ in the US.
The accolade by Startup50 is the latest in a string of awards for the three-year-old Silicon Valley company that has close links to Australia, its CEO and head of sales both based here.
Saisei founder, and now chief strategy officer, Bill Beckett, told Computerworld Australia that the tie-up with Kestelman came as a result of his work for Dodo ahead of the sale of the company to M2 Telecommunications (M2 bought Dodo in March 2013 leaving Kestelman with about $100 million in cash and $50m in M2 shares).
“Larry found out what I was doing to develop the product [Saisei]. One day he called me out of the blue and told me he wanted to invest in the company and now he is our largest investor and a great believer in our product,” Beckett said.
Saisei announced in February 2014 that it had closed a US$5.6m Series A round of funding that was led by Kestelman’s Oxygen Ventures and included participation from several early seed investors.
Kestelman said at the time: “The patented technology that Saisei has developed is truly unique and has the potential to revolutionise the way network operators see and control their networks.”
The Dodo founder is now a director of Saisei. Saisei’s Australia-based CEO, Boris Rozenvasser, was formerly consumer director at M2 Group. Melbourne-based Julia van Heerden, Dodo’s former marketing manager and later GM of consumer marketing at M2, has an advisory COO role at Saisei.
Collects and analyses all network data
According to Beckett, Saisei’s main differentiator from other network analytics tools is its ability to gather complete data on traffic to enable very granular monitoring and control of network behaviour without imposing a huge performance overhead or requiring large investment in data analytics capability.
“We differentiate ourselves from any other vendor in the world in that we can do analysis right at the edge,” Beckett said.
“The number one value we provide for customers is a network analytics sensor that can go into a network, look at the data and automatically start doing analysis on millions of flows across the network in real time and give them what we call ‘big data in a little box’,” he said.
“When you plug in our sensor, it starts collection millions of records every second and building user tables and application tables but doesn’t stream it out to a server. Our competitors’ technologies have sensors that sample the network — typically one in 2000 packets — and send those records to a collector to crunch it, so at best you get only sample data.”
“We can analyse 100 per cent of very session we see across a wire, whether it is one gigabit or multi-ten-gigabit,” Beckett said. “We can scale to millions of flows. We analyse these flows by 45 metrics and we maintain this table completely in real time for every packet in every flow using our analytic engines.”
The sensor can function either as a passive monitor, or can be used to control network behaviour based on analysis of the information it gathers.
“What makes Saisei unique is that it can use the analytic data to enforce policy or to make the network run better,” Beckett said. “An example would be fair usage. In real time it detects how many users are active on the network, how many sessions they have and what type of user they are, so that each one gets their proportionate capacity.”
He added: “We have a big customer in Australia that had an application that was not behaving very well. Once or twice per day it would act up and consume much all the network capacity.
“A normal response to that with other technologies would be to try and pinpoint the source of that traffic, maybe by the IP address or application type. What we are able to do is to set a general analytic policy that says ‘If I see an application that exceeds x percent of my capacity, quarantine that application so it does not interfere with the rest of the traffic.”
Role in SDN and NFV
Saisei also sees a role for its technology to support the development of software-defined networks and network functions virtualization. The company joined the Alcatel-Lucent CloudBand Ecosystem and became a member of the Open Networking Foundation in June 2014. Beckett said that organisations migrating to SDN and NFV would need the greater real-time visibility and control over applications, users, and locations that Saisei could provide.
The company now had dozens of customers across seven countries and will be selling in more countries shortly, according to Beckett.
“A lot of our customers are every recognisable brand names. In Australia we have some large retailers using our product.”
Known Saisei customers include Sunway Digital Wave, a managed services provider to Malaysia’s largest property-construction group; First Cagayan Converge Data Center in the Philippines; NHN Entertainment, Korea’s third largest game company; and Fort Bend County Libraries, Texas.
Beckett said Saisei aimed to sell primarily though channels and by partnering with other companies with complimentary technologies.
“We do a combination of direct and channel, but mostly channel. We don't want to build large sales organisations around the world, and there are a lot of companies providing visualisation tools that what want to partner with use to go to market.”