December 2, 2019
By Tim Siglin Contributing Editor, Steve Nathans-Kelly
Featured Articles

SMW '19: Tulix Talks Low-Latency Streaming

Read the complete transcript of this interview:

Tim Siglin: Welcome back to close to the end of the first day of Streaming Media West 2019. I'm Tim Siglin, Contributing Editor of Streaming Media Magazine and the founding Executive Director of the not-for-profit Help Me! Stream. And today I've got George with me, who is one of the founders of Tulix. George, for those in the audience who may not know, what is Tulix?

George Bokuchava: Tulix is a complete software streaming solution company at Atlanta, Georgia. We've been in business for many many years. Many years, I'm talking about twenty plus years.

Tim Siglin: Right.

George Bokuchava: And our foundation is knowledge of the internet. We were actually one of the pioneers of internet many years ago. And what we decided to do about 10 years ago is to focus at streaming.

Tim Siglin: Okay.

George Bokuchava: Streaming services. And our background and knowledge and experience of the internet gives us a huge advantage in building solutions which are robust, which are cost-effective and which are user-friendly. So if we can say about circle.com, you probably remember .com era where companies were jumping in into the internet?

Tim Siglin: H-hmm.

George Bokuchava: We are coming from that era. And we are one of the survivors who actually, at this point utilizing all our experience and all our knowledge, infrastructures that we build. People are staying with us for twenty years or more. And this is how... managing our company. One of the features maybe, or something to mention is we have twenty-plus-year-old history. We have some customers who are staying with us for more than fifteen years. And probably there is a reason why.

Tim Siglin: When you say an end-to-end solution, that includes: encoding, trans-coding, content delivery player support?

George Bokuchava: Absolutely. We have a very interesting approach. And I don't know about any other companies probably do exist. But we have more interesting approach. We have a so-called default setup. Default setup, we're using a number of third-party products which are very cost effective and they have sufficient quality to satisfy probably eighty percent of the customers. Now, there's some cases where we have very specific requirements. And in those cases, we'll be happy or we'll be upgrading any components. Would be it encoding, would it be CDN, would be any other parts of the complete OGD solution. Would be in just, we're upgrading with any solutions which are best in the market. And we're talking about scaling from the smallest and the most inexpensive solutions to the most expensive and has the best solutions available to market. Whatever fits the customer.

Tim Siglin: So one of the things I know you and I have talked about in the past is how important latency is. And it seems like the industry is finally, after going through a number of years of, let's scale by things like fragments, and chunking, which inherently adds latency to the delivery, are now really concerned in the opposite direction that we don't have low-latency streaming. You've been around for twenty years as you say, is the push for latency something new? Or is it something that we painted ourselves in the corner with fragmented solutions that we're trying out how to get out of that corner and back to lower-latency delivery?

George Bokuchava: It's interesting. It's also a very big question or subject to talk about. The interesting part, before chunked video solutions, like HLS, DASH, CMAF, and all those stuff there's also so-called RTMP, RTSP. Or actually EMBONE, when you talk about multicast was over-the-internet solutions. We're talking about almost no latency or so any significant latency that no one was talking about. Now, because of the convenience of CDNs, because of some specific hardware products. Like Apple, for example, there'll be one of them. There was so-called HTTP where chunk-based protocol was pushed. And convenience is great. Actually cost-savings are great for CDNs which obviously is transferred to the customers. But then suddenly the problems with low latency showed up. It's a true problem because not only latency from the point of you compare to the original stream. But there's also problem than you have seen Wise is in the same room and all this time we're playing different--

Tim Siglin: No synchronization.

George Bokuchava: Yeah, no synchronization. Imagine if it's even on the engineering side or monitoring side, it could create difficulties. So, the definite multiple solution for us actually, we're in the very good position because with an infrastructure that we have with optimize routing out of just couple of bigger notes that we have, we can push any protocol. That's any protocol. It could be anything, actually, including for their correction protocol between us and the clients or the players who can do that too. But obviously we're about to see for us, it took half a day for us to implement. We're about to see as a demo for one the shows.

Tim Siglin: Sure.

George Bokuchava: But on the other hand, we have obviously partners and partners who are pushing us to get a latency solution. But my response was, let's wait for the biggest guys. Those biggest guys will come up with a solution. That solution probably will become a standard. Now, no one knows for sure, but I think one of the recent activities were done by Apple. They came up with lower latency.

Tim Siglin: Right.

George Bokuchava: HTTP or HLS protocol but solution. And I think there could be a good chance that there could be a good solution. And if not that, something else, in my opinion will show up which will be pushed by big players. That way we'll not have to have all the small solutions. Like, we have in our company five different solutions, protocols. And all of them are proprietary. You can name, but you know, it's really five different solutions. I'd rather wait until the mains will show up and fully support.

Tim Siglin: And you mentioned WebRTC, of course, just like RTMP, WebRTC has a scaling issue, without using something like content delivering network. So where is the balance between very low latencies like WebRTC and very scalable solutions like HLS and DASH? Do you sense there is a middle ground between those things? Or will the push be to get closer and closer to WebRTC latencies? Rather than the multiple second col lase we've had with the chunking solutions?

George Bokuchava: In my opinion, we're about to see bigger problems. Just my opinion. Than just scaleability. Due to the nature of our activities, we're dealing with a lot of different platforms.

Tim Siglin: Okay.

George Bokuchava: Starting from smart TVs and going through all these Amazon Fire, Apple TV, and you know. almost any device which is capable of playing. And in many cases we have conditional access, in many cases we have digital rights management, And you know it's easy to come up with an almost lab-based solution where you get a stream on one device or just couple devices. And kind of clean... But once you go into full production mode, then you have to support like even, I don't know, Hollywood studio stuff. I can't imagine what we're about to see at this point, being sufficient or utilize for any solutions.

Tim Siglin: For major live sporting events, where piracy is a big concern and authentication's a big concern.

George Bokuchava: So to answer a little bit closer to the response that I think which will be a chunk-based protocol, which will be modified or it will be tuned. Not just by decreasing the size of the channel but whatever technology will come up. The way in there is to make it as close as possible to real time. Or potentially it could be a completely new protocol but it will be, in my opinion, much more painful to implement it. [Tim Laughs] Because of our variety of do I see

Tim Siglin: Sure.

George Bokuchava: Again, you can make anything you want. Let's say I have maybe five different protocols of which initially one of them is done by a company in Atlanta where we're operating. They have their solution which being used for one of the largest auction houses in the world.

Tim Siglin: Okay.

George Bokuchava: But again, it's just one solution. Going mainstream with the solution, my opinion is a different story.

Tim Siglin: All right. George, thank you as always for your time. We'll be right back with probably with one more interview before the end of the day.

George Bokuchava: Thank you very much.