November 23, 2018
By Tim Siglin Contributing Editor, Steve Nathans-Kelly
Featured Articles

SMW '18: Mainstreaming's Mike Smith Talks Hypernode Technology

Learn more about content delivery at Streaming Media's next event.

Read the complete transcript of this interview:

Tim Siglin: Welcome back to the second day of Streaming Media West 2018. I'm Tim Siglin, contributing editor with Streaming Media Magazine and the founding executive director of the not-for-profit Help! Me Stream.

Tim Siglin: Today I've got our first interview of the day with Mike Smith from Mainstreaming. Mike, you're a VP of engineering, is that correct?

Mike Smith: I am, yes.

Tim Siglin: What is Mainstreaming?

Mike Smith: Mainstreaming is a brand new content delivery network, one of the newest ones in the industry. We are unlike a lot of the other CDNs who have had a foothold for some time now. We're very video focused, very video-centric.

Tim Siglin: Interesting. And is that live video, on-demand video, or both?

Mike Smith: Live and on-demand video, yes. And audio as well.

Tim Siglin: Okay. From a live video standpoint, what are the unique selling propositions for you all? You say you're video-focused, but what is it that makes the live with you better than, say, the live with Akamai or Limelight or somebody else?

Mike Smith: That's probably the most common question that we get, is how are you any different from the Akamais and the Limelights of the world, that have been at it for a while? As I mentioned, we're very video focused. But we've architected our network for very different delivery of pieces, elements, images, whatever it may be. Over the years, they've been able to adapt that technology and stream video fairly efficiently and fairly effectively. Where we came at this is, ultimately, a little later in the game than some of the other CDNs; however, we focused entirely on video.

Mike Smith: So we built what we call our hypernode technology. It's really, using a marketing or sales term, it's the special sauce.

Tim Siglin: Okay.

Mike Smith: Essentially, our hypernode technology is an orchestrator. It's an algorithm that can make intelligent decisions, can dynamically route and change pathing on the fly to create a very smooth and very high-quality experience.

Tim Siglin: And will that also work from a load-balancing standpoint?

Mike Smith: Absolutely. And where we've leveraged our network differently than some of the other competitors is, rather than ... For many years in the CDN space, it's been a numbers game. How many POPs, how many servers do you have? For many, many years that was a very relevant question. As we deal with the topography of the public internet today, it's very different. So what we have focused on really, is paid peering arrangements with different interconnections and ISPs.

Tim Siglin: Good, in fact I was gonna ask about that, given the fact that we went through the whole federated model for a while and that didn't really pan out. Peering arrangements seem to be a key element.

Mike Smith: They really are. That's actually given us a lot of ability. Today we operate the sixth largest interconnected network in the world, actually.

Tim Siglin: Oh wow!

Mike Smith: Yeah, surprisingly. So we've been able to, to make up ... That's really been from a foundational level, any CDN is only as good as its foundation. So we've determined at a foundational level, that's really where we had to really put a finite focus in, is doing that.

Tim Siglin: Okay, that makes sense.

Mike Smith: So, 2500+ interconnections globally so we're able to do more with less.

Tim Siglin: Okay.

Mike Smith: Our goal, ultimately, as a CDN, is to create what we call our one-hop experience. Only one hop traversing on the public internet. And that really allows us a lot of economies of scale. Of course, it allows us to operate a bit leaner than some of the other folks.

Tim Siglin: And that one-hop includes the last mile in to the consumer?

Mike Smith: It does, yeah.

Tim Siglin: Wow!

Mike Smith: Our goal is to get from the edge of our network to the end user, the end user's demarc if we can, with one hop.

Tim Siglin: Okay, very nice. Now, obviously, from a latency standpoint, if you can achieve that goal that's pretty significant. So it sets you up nicely for live. On-demand though, having the 2,500 that you described, does that mean you have to have a plethora of storage and caches and all those edge locations?

Mike Smith: No, it's a very good question. We're able to leverage a couple of main distribution points for storage per se. Ultimately, we've been able to cross-connect and interconnect our network in such a way that we're able to operate very efficiently with the interconnections the way they are today.

Tim Siglin: Mm-hmm.

Mike Smith: On-demand, we support a number of different models. So, today with our dynamic ingest abilities, our hypernode technology, I should say, isn't limited to just edge-based delivery to the customer. We also use it internally. So from our ingest to the origin of the customer, allowing for a dynamic mesh if you will of ingest.

Mike Smith: So, that allows us to operate very efficiently as well. So in the sense of and with a lot of mid and large size broadcasters and a lot of our current customer base today, we have quite a blend of customers who will want us to rely on them as the origin, versus our smaller to mid size customers that want to essentially upload or use our APIs to put items into our storage.

Tim Siglin: Okay, interesting.

Mike Smith: So both models, yeah.

Tim Siglin: In some ways you have a hybrid between what the traditional Akamai, multiple POPs and storage at the edge, and Limelight which is strictly just a core network.

Mike Smith: Absolutely.

Tim Siglin: That's interesting that you're taking best of both worlds. Do you also support legacy streaming formats?

Mike Smith: Icecast, Audiosuite, and Shoutcast Audiosuite.

Tim Siglin: Sure, sure.

Mike Smith: While not as common in the video space, there's still a lot of the radio broadcasters out there that are leveraging that technology and we still fully support that.

Tim Siglin: If I were to sit down with you at Streaming Media East in May of 2019, what would be the trends that you would've seen between now and then?

Mike Smith: That's a good question. We have seen such an explosion of OTT and ultimately one of our focuses, we have many different focuses and industries that we focus on, however, one of them being of course OTT broadcasters, even faith-based ministries.

Tim Siglin: Sure.

Mike Smith: And we've seen a huge push in not only adopting that technology, but education folks. The broadcast space, for so many years has operated in, in maybe the term in a vacuum is a bit extreme, but in their bubble. Iron transmitters or steel transmitters at the top of mountaintops and very traditional broadcast elements. So the space of OTT initially broadcasters didn't want to wrap their arms around that. They saw that almost as, ultimately the enemy.

Tim Siglin: Very true.

Mike Smith: Very true, initially. We've seen such a shift in recent years of the traditional broadcaster really wrapping their arms around that and saying oh yeah, I need to be an OTT and in fact, I can get to more people via OTT than I could through traditional means.

Tim Siglin: And that brings up an interesting question. One of the discussions I had yesterday with an interviewee was around the whole, you know, traditional engineer, class 1 license, versus the IT guy. The IT guy isn't going to go get a Class 1 license to be qualified as a broadcast engineer. Broadcast engineer may be a little hesitant on packet movement, but ultimately understands that IP delivery is coming and that's where we need to focus. Where do you find the balance of people who understand both, or is that still a dichotomy?

Mike Smith: Yeah, it's interesting. So, initially in my career, I was actually a broadcast engineer. I started off as a radio engineer. Went up to transmitter sites in the tops of mountains and worked on analog gear back in the day. So, I think initially as, I like to refer to them as the old graybeards.

Tim Siglin: Yeah, sure, sure. I mean, you don't have gray in your beard, but...

Mike Smith: Oh stop it, I have a little. No, I mean the old graybeards that are out there, ya know, initially they were very resistant. And knowing that community very well, being plugged in with the society of broadcast engineers...

Tim Siglin: The tests they were having to take to be certified had nothing to do with IP delivery itself.

Mike Smith: Exactly, it didn't. Not at all. So initially they were very against it. I think once that subset of folks embraced the fact that holy mackerel, I can do a heck of a lot with IP, we saw that in the radio business where companies like Tieline for example, came out with IP enabled gear allowing you to send broadcast quality stereo audio at CD sample rates over the internet and with very low latency.

Tim Siglin: To replace the zephyrs and ISDN tie-downs.

Mike Smith: You got it, the old zephyr ... the expensive ISDN, bonded ISDNs for zephyrs and such and even some of the old POTS gear to transmit audio.

Tim Siglin: Yeah, yeah, absolutely.

Mike Smith: When we saw that subset of folks really start to embrace that, I really feel like the rest of that kind of fell into place.

Tim Siglin: Okay.

Mike Smith: Because as you look at kind of the modern-day newsroom or the modern-day media organization, it's made up of a very strong subset of everyone from millennials, yes I'm using the magic word to x-ennials, to even the gen-xers and that generation of working, of the working staff in these organizations, they're in social media, they know the value of social media restreaming. They know the value of delivering online. And I think there is, I think it's accepted now and it will continue if you ask me the question at East I would say this, that we continue to see the value of delivering content on the internet. Whether that's IP-based TV like YouTube TV and so being in situations like Mainstreaming and being very video focused on this, ya know, we see tremendous growth. I mean billions and billions of dollars of growth.

Tim Siglin: Sure.

Mike Smith: So, we plan to stay the course and continue to build out our network in the US and then continue to expand our US operations.

Tim Siglin: Well, from an engineering standpoint, I'd say Mainstreaming is lucky to have you having that background and this background. We'll be right back. This has been Mike Smith from Mainstreaming.