Bonus To Boldly Go Where No #VoIP #Collaboration Has Gone Before: Subspace
SPAM was one of my first lessons on the power of the cloud. In the 90s, the Internet was still too dangerous for the enterprise. I worked at one company that required a note from the CISO as well as a dedicated, off-the-corporate-LAN PC to get Internet access. The killer app of the Internet that changed all this was email, and it blew-up in the in the early 2000s. Though email was old, it became became mainstream for business communications in the 2000s. But, there was a problem, and it was called SPAM. Unsolicited and dangerous junk mail was steadily increasing. There were several times that I assumed email was doomed – that SPAM and junk would make the medium useless.
IT administrators were fighting SPAM by installing anti-SPAM solutions on their mail servers. Though more primitive, the threats back then were much more dangerous than they are today. Anti-SPAMware scanned for known source addresses, known test, and mostly known malware attachments. I vividly remember the Melissa and I Love You attacks that crippled enterprise email globally. As Microsoft hardened Outlook and Office for malware, the SPAM threats evolved to phishing attacks and other lures to trick employees to voluntarily give away information. This is how many attacks occur today. These emails are much harder to catch with a traditional SPAM filter.
The solution wasn’t a better software software for the server. I’m not sure how nailed it, but I experienced this on Gmail. Google approached SPAM mitigation through filters enhanced by crowdsourcing. Quite literally, its millions of users were asked to flag SPAM. Today we call this AI training, but crowdsourcing was the magic ingredient. Google expanded the concept to Maps, and also acquired Waze because it did it even better. The concept is to use actual data, not libraries to make routing decisions. There’s gold in them there network nodes.
That’s the story of Subspace. But it’s not limiting SPAM, it’s optimizing network routing. And it’s data doesn’t rely on users voluntarily flagging bottlenecks, it harnesses real global network performance – specifically related to games. It’s a brilliant evolution that should be particularly fascinating for anyone involved in real-time communications.
Talkingheadz interview podcasts are available on most podcast directories.
Dave Michels 0:00
Hey, everyone, today we are going boldly where no comms podcast has ever gone before into subspace.
Evan Kirstel 0:20
So the guest is Shatner, William Shatner.
Dave Michels 0:24
No, no, no Don’t we did we didn’t get Shatner? I think he’s actually retired these days. But we did get mo nezzer. Roddy, the president of voice I want I want that title president of voice from sunspace.
TalkingHeadz Is a semi monthly podcast with interviews of the top movers and shakers and enterprise communications and collaboration. Your hosts are Dave Michaels and Evan Kirkstall, both of which offer extraordinary services including research, analysis and social media marketing. You can find them on Twitter, LinkedIn, or at talking points calm. That’s points with the Z and Devin curse go calm. That’s kr STL.
Evan Kirstel 1:05
Welcome Mo. Hey, Dave. Evan, great to be with you guys today. Thanks. So you know, I guess the obvious first question is, what is subspace?
Mo Nezarati 1:17
Well, subspace is the place where real time application traffic moves the fastest back boldly, when starships needed to communicate with each other over very long distances, they use subspace while here on Earth, we haven’t gotten to those starship times yet. So in order for traffic to move really well, from on real time applications, we’ve now built this environment called subspace where traffic moves from one point to the other, across our hundreds of pops around the world, using our special software that calculates and analyzes where traffic is being bottlenecked and stopped. And we ensure that all communications goes from a to b, the fastest possible way.
Dave Michels 1:59
So I just wanted to be clear, because this is an audio podcast that Evan is the one wearing the red shirt. I don’t want any, any confusion. There’s any aliens listening to this, you got to help me out here. Because talking heads out this podcast, this is all about enterprise communications, I think a subspaces games or whatever, we cover serious stuff like Solitaire, but that’s what’s coming anyways, but a good podcast, and we’ll invite you back when you get serious.
Mo Nezarati 2:26
Well, it’s interesting before COVID, when we first started, we we were solving problems for multiplayer gamers who were sitting at home, connecting to remote servers where their games were being played. And guess what a lot of us now are sitting at home connecting to remote infrastructure in order to get our work done. So there is actually a connection between how we solve problems for people sitting at home, connecting to those remote infrastructures for games, to now people doing the same thing within enterprises. So the way that we’ve all worked on the evolution of how we work, this changed, and subspaces, I think now a perfect solution for these enterprise customers who used to have like 20 offices that they connected, and now they’ve got 20,000 offices that they need to connect so so we solve all kinds of problems for people.
Dave Michels 3:12
So it’s all fun and games until your internet goes out. That’s it exactly.
Evan Kirstel 3:16
So I get what you’re saying my teenage son is a gamer and when he’s on the latest fortnight load that completely destroys my home internet. So clearly games and enterprise collaboration happen over networks, you say you’re a collaboration game for the enterprise at subspace are what are you doing exactly?
Mo Nezarati 3:38
Well, not quite a collaboration game, you can think about us as a parallel speed of light network that is designed specifically to solve the movement of real time traffic. In the last 10 1520 years, the internet has really evolved in terms of providing solutions for volumetric data, things like Netflix, YouTube, the latest download of your iOS software that you get when all of a sudden a billion people decide to upgrade. So all of those sort of stored images of information are being pulled across the internet to people. And it feels like wow, the internet has gotten faster for us. But in fact, it’s only gotten better for static volumetric data hasn’t gotten better for real time traffic. So what we do is we make make that movement of real time traffic a lot better and faster.
Dave Michels 4:30
I’m looking at your internet site, and then he got the tagline that you’re taking control of the internet, like that’s the other. That’s the outer limits reference. We now return control the internet to you. So is the internet out of control? I guess the first question and isn’t that kind of what I know of as network acceleration, and that’s no game. Isn’t that kind of a crowded sector for a startup?
Mo Nezarati 4:52
It’s not so much a crowded sector for startups right now. They’re, like I said, there’s a lot of solutions in and around the He’s sort of the content delivery networks that solve these kinds of problems. But there is very few companies and you know, other than sort of SD when solving problems for off of remote offices connecting to one another, no one is actually paying any attention to accelerating traffic from infrastructure, all the way down to someone’s home, and, you know, wherever they may be on from any point in the world to any point in the world. And I think the way that we’ve deployed our software and hardware stack, really is a massive differentiator in in solving some of these issues.
Evan Kirstel 5:33
Okay, no, I think I get it I in my previous life, I was an XPS SBC expert at Acme packet, the inventor of session border control, I learned a lot about the edge, but you aren’t an SBC vendor. So I guess I don’t have to buy a $90,000 box to put in my home. So what is your edge? Exactly?
Mo Nezarati 5:57
Our software is really precise in calculating network paths. So some of our secret sauces, in terms of how we solve these problems is I like rock sauce. We broadcast the same IP address around the world through our pops to a destination. Let’s take an example of a user and an enterprise space connecting over a Citrix session or a virtual desktop session from one end to the other. They typically put in the IP address or you know, a DNS name for that Citrix server, and then they’ll connect over the public Internet to that destination. So what we do is, we broadcast out a proxy, a replacement IP address and port for that server that they’re trying to connect to. And we push that same IP address across all of our pops. So we will be five milliseconds away from the user sitting in Mumbai, India, and we’re five milliseconds away from the user sitting in in Sydney, Australia, or Frankfurt, Germany, or, you know, Denver, Colorado. So as those users join that IP address from anywhere they’re located, we carry that traffic all the way to the destination. And then our software calculates very precisely what all the possible paths are, from the Ingress point on the network to the egress point where we need to get out and reach that destination. So those calculations are done, like on a millisecond by millisecond basis, and we connect with hundreds of ISVs around the world, and we calculate exactly which paths are the fastest paths. And we continually calculate those paths and and the speed at which we need to deliver those packets. So at any given point, if we if we pick a path for you, that is slower than the one that is available, you know, five milliseconds later, what we’ll do is we’ll actually move you on to the next path. So we’re continually measuring an inflight, we can move your transmission of your packets from one path to the other, we weather map how internet traffic is moving, because of all these, like millions of gamers that are on our network, playing their games. And at any time, if there’s any congestion, there’s wire cuts or anything else, we essentially we can detect that and move the traffic over to a different path without any interruption. So it’s a really interesting type of solution.
Evan Kirstel 8:16
Wow, sounds almost magical. So are you looking at traditional corporate networks, or is this more about the last mile to the home or both?
Mo Nezarati 8:27
It could be both what we have customers that are looking at connecting their offices through us there, you know, located in far off locations, they typically have relied on MPLS links, or you know, SD wins in order to connect these offices together. But the problem is when you have now as I mentioned, if you have 1000s of people working inside of their homes, that you now have 1000s of offices, wherever you might, you may have had 10s of offices in the past. So corporations need to now sort out how to connect and make all of the remote home offices as part of their collective see seamlessly. But we also have people coming and going in any environment. So it does not want to necessarily come up with ways or solve for, you know, shipping hardware and configuration and helping onboard and off board, these hundreds of people that might join your company on a given month. So our solution not only can solve the problem with physical office locations, but it also extends to all the remote offices that people now have with their new work from home realities. So we deal with both the corporate networks as well as the last mile. And we always do all of our calculations all the way down to the edge. Because we work with and we peer with hundreds of ISP around the world. Each one of our pops is connected to all the best providers in that region. And we can move traffic natively from one network to another network across are seamlessly without having a lot of bounce overs between these networks.
Dave Michels 9:57
I’ve never really thought about working from home as kind of a Again, but I guess it’s just a different game just not a very fun game. It’s, I guess it’s a rat race gamers like that. But when you’re working from home, is there any kind of particular workload that you are helping optimize? Is it voice, video text, or what is subspace optimize?
Mo Nezarati 10:17
What we fundamentally optimize the the network connections from any client that they’re running at that time to the server that they’re connecting to. Now, those servers could be video conferencing servers, like your zoom server, your Microsoft team’s server, Your Cisco WebEx server, etc, we could be accelerating the paths there and making sure that your connection is stable and reliable all the way through. But it could also be, you know, a contact center agents who are receiving inbound calls, sitting at home connecting to contact center infrastructure that’s far, far away from them. And as you guys know, in this space, you’ve got a lot of experience, you’ve got agents that may be sitting in dozens of different countries connecting to the same infrastructure. So because of the nature of how we, we’ve deployed our software, and the fact that we broadcast the same IP address all around, we are so close to most of the population are where those agents are sitting. And, and we can carry that network traffic. Lastly, we’re seeing a lot of demand for virtual desktops. Again, in a lot of verticals like the finance, vertical, virtual desktops are used a lot, because you don’t want people storing, you know, financial information that onto their computers, everything sort of done through thin compute. And we accelerate those virtual desktop environments. But we’re also seeing solutions like AI assist solutions that are looking for better latency how these products work, essentially, is they analyze the conversation between you know, two parties, and then they try to say, hey, did Evan have a similar problem or with somebody else before or Yeah, Dave, solve that problem with a different way. So those AI assist looks at the knowledge base to try to identify solutions. So the timeliness of identifying those solutions, picking up the keywords, and then responding back to the agent is really important. So if we can start shaving hundreds of milliseconds and seconds off of that transaction time, when the agent is receiving those AI assist suggestions, becomes much more relevant to the conversation they’re having. So before they move off, on off on to the next topic. So we’re seeing all kinds of really, really interesting applications.
Dave Michels 12:19
This is really intriguing, because, you know, we talk a lot about work from home and all the transitions that have occurred, but we haven’t really talked that much about the network. I’m glad you talked about the agent at home. And that’s the one part of enterprise communications, the the contact center, that we that’s actually named after a physical place and not anything else. Do you think that agents at home are going to last? Do you think they’re gonna be heading back into the office at some point,
Mo Nezarati 12:42
it’s hard to say how that’s going to change in the future. But regardless, you know, I look at sort of the history of real time applications over the course of the last 10 years or so, if you look at, you know, when we have 3g networks on cellular, and I think the, you know, cellular milestones are interesting ones because I
Dave Michels 12:58
love the beegees man, the beaches of great,
Evan Kirstel 13:01
Dave, Dave still has 3g, he’s really hooked. For sure. So
Mo Nezarati 13:08
that’s, that’s the whole staying alive. And, and being connected is what all of these applications, real time solutions are all about. But you know, when we have 3g networks, you had only voice applications, potentially, that were working as soon as we had LTE come in. And with 4g, we had Tick Tock and, you know, video applications and Instagram and all sorts of new applications that that popped in. Now, with 5g rolling out, I think there’s a whole new layer of applications that will come in, and a lot of these new systems are going to rely on real time networks, you think, you know, we’ve all read about machine to machine communications, IoT sensors everywhere. If the information that these sensors are sensing are not sent back to environments where they’re being processed in a timely manner, they, their relevance becomes less so I’m not really worried about whether this lasts or not, I think we’re in it for a little while anyway. But regardless of whether agents stay at home or come back into a physical delivery center, I think we’re going to see it, you know, the number of real time applications increase. And I think you we are at least underpinning our company on solving this problem, in a broad sense, much more dynamically going forward.
Evan Kirstel 14:18
Exciting. So contact center agents are all about service. And often mission critical to the company’s operation is sunspace able to provide that premium customer service, whether it’s availability or reliability or speed, network, QoS, etc. for that for that mission. Critical operation,
Mo Nezarati 14:40
for sure. I mean, we’ve talked a lot about sort of subspaces ability to give performance all the way down to the edge. I mean, if we look at how it managed delivery centers in the past, they manage the connection, they manage the performance, they manage the security of these locations, well with agents going home, they’re no longer able to guarantee the kind of performance that agents are going to get all the way down to the edge. So the beauty of subspaces, we give that performance all the way to the edge. But we really simplify the deployment. Because we don’t require any hardware or software to be installed at the edge or on the server, we are a pure network solution. And we do all of our problem solving using IP addresses. So the elegance and the simplicity of the product itself is really part of our magic in terms of like solving these, these problems for it without really burdening them with any kind of solution that is going to that involve more training or more it involvement. You know, if we look back at the example I provided on the on the remote desktop, you’re literally replacing an IP address with a different IP address, and that IP address is broadened across broadcast around the world. So we’re really simplifying, in fact, how people connect and and and how people work.
Dave Michels 15:51
And I gotta call you out on that Mo. I know, nobody comes on our podcast, and guests explain things away with magic, can you give us an idea of some of the technology you’re actually using?
Mo Nezarati 16:03
For sure. So we literally are monitoring how traffic is moving. And we’re weather mapping in real time, we have precision timers, that basically measure traffic and how data is moving across the internet in nanoseconds. So our timers are incredibly precise. And we use technologies that maps out a source to a destination in millions of different paths. And we’re always comparing these paths to one another in real time. So we have,
Dave Michels 16:32
like ways we are
Mo Nezarati 16:35
exactly so we’re in fact, all of the traffic that moves on the weather mapping that we’re doing is similar to how ways or Google Maps and others are looking at traffic analysis on a physical road, given the information they’re getting from people’s phones sitting in their pockets. So we measure that traffic and we want
Dave Michels 16:54
to help you out with your metaphors you you want to do in weather stuff on ways and you do things in flight. You too many metaphors here. I’m trying to dumb it down for you, Dave. Well, let’s
Evan Kirstel 17:08
discussion, let’s shift the discussion from the theoretical, abstract metaphors to customers. Tell us what traction you’re getting in the market with real customers and developers.
Mo Nezarati 17:22
We are getting great traction. We have tons and tons of developers, we already have a waiting list of over 10,000 developers that are waiting to onboard our service. We have enterprise customers that are clamoring to get on various different services that we’ve got whether it’s our you know, sip, teleport product that accelerates sip traffic, or our global term product, which is a total disruptor in the whole web RTC space, we have customers that are looking to get onboard and start solving this and to the point I’ve made earlier is we are fairly unique in the way that we’re solving this problem. And that’s why we have so much interest from customers already.
Dave Michels 18:01
Do you have a channel? How do I how do I buy some space? I’m just I’m guessing I don’t go over to GameStop?
Mo Nezarati 18:08
No, we’re hoping to be a hotter stock than GameStop GameStop. In a few years now. We’re building our channels to market right now. Our developers can basically, you know, head over to subspace COMM And there’s a sign up button right there and they can sign up and join the party.
Dave Michels 18:24
When you want to be the opposite of GameStop. You want to be game go? Yes,
Mo Nezarati 18:28
exactly real time love it and going so fast that you can see us move.
Evan Kirstel 18:32
And more practically speaking, it does seem logical to go through UCAS and seek as providers is that your road to market as partnering with with those folks and having them resell or bundle subspace?
Mo Nezarati 18:46
Absolutely, we are in discussions right now with a number of eucast and C cast providers. And the way we’re going to market with them as a solving problems individually for some of their larger clients. So if you’re a DPO that has you know, 10s of 1000s of contact center agencies and you’re looking for a solution and you have a provider. We’re working with them to solve the problems for those individual companies. But we’re also working with them to upsell and bundle subspace into the existing offer. So if you have a voice or a video platform like many of the eucast providers have, if you’re looking for a better experience for your users, we can bundle subspace and make those experiences better. It really stops the need for those providers to start to increase their infrastructure footprint. Because we can bring traffic and users to them faster that they can build out infrastructure and connect all of these systems together.
Dave Michels 19:37
That makes a lot of sense that you would go through the providers that you guys are seek as providers. Some of them have already got a kind of gotten into network services with SDN offers. How is this different than say SD win?
Mo Nezarati 19:52
Well, SD win is great for you know, fixed points and it does it does a great job at that. I think we complements you know companies investment in SD win by really taking traffic all the way down to the edge. While we have some customers who want to use this exclusively and eliminate some of their MPLS, or SD win infrastructure, many of our customers are using us first, for their remote workforce where SD Wan is not really a practical solution. And from there, you know, we can sort of explore how we go forward. Generally, you know, with we think that we’re sort of like the next generation of SDN and, and we have features that we’re building into the system that allows you to sort of compose your own network, put in rules around who joins from where, which directions, and which countries to, you know, have a network travel through which countries to avoid, etc. So we have all kinds of really interesting solutions and composability of the network that I think makes our solution very attractive for companies that have specific compliance needs or other other issues.
Evan Kirstel 20:54
And clearly subspace has benefited from the pandemic, I actually can’t think of another startup that has better timing than subspace given the reality of customers needs right now. But what what comes next post pandemic knock on wood? How do you see your customer landscape changing, if at all,
Mo Nezarati 21:14
I think the opportunities as I mentioned, you know, they’re they’re boundless with respect to the need for real time networks. So I see the need increasing, whether it’s, you know, solutions, like AR VR that require real time feedback from users and environments, whether it’s IoT, you know, speed has always been a premium. And when, you know, companies have always wanted to get a better, faster, safer internet. Now that we can provide that safety and security and performance all the way down to the home, it really gives people an opportunity, for example, at subspace, we’ve been growing our employee base, but instead of looking locally, where we would have done in the past, we’re now hiring the best people who can do the job, no matter where they’re located. And the ability for us to sort of extend our offices to anywhere in the world, you know, just by connecting people together, really, I think extends every company’s ability to manage and grow their footprint and their customer base. So fundamentally, I think that the need for real time traffic and real time solutions is on an upswing, and we’re seeing requests from, you know, database synchronization to IoT, we have, you know, we actually have like a university right now that has a simple problem with like, their, their, their users with SSH, the performance of SSH, so we’re just simply solving that for them. So love, it’s love. It’s, it’s amazing the amount of opportunities that I think exists that telemedicine is coming along, you know, so there’s all kinds of very interesting solutions that people are building and we haven’t even dreamt up the next generation of products. You know, when we have 3g, we didn’t dream of solutions like Tick tock, you know, what will 5g Solutions look like? As we have better networking across devices?
Evan Kirstel 22:59
Well, please connect Dave Michaels to subspace because I’ve been trying to get him off of that fax machine for years. So subspace might be, you know, an excellent opportunity to finally get the facts, Dave. All right. All right.
Mo Nezarati 23:14
I sold him that fax machine. So you know,
Dave Michels 23:18
you mentioned a little history there because we can’t we can’t let you go without asking you a little bit about little history. We just had the visa and Laura on from Avaya just a few episodes ago, and they were talking about you’ve probably never heard of this Avaya spaces. But I believe you were one of the master architects of its prior previous name, Zane. I was kind of surprised to see you left. Avaya was one of those. If you love something, you’ve let it go is that time was up the logic there.
Mo Nezarati 23:45
I really enjoyed my time at Avaya. I learned a lot from the people there. And the team. I think, you know, when I had when I joined the company after selling esna to Avaya. In 2015, I had committed to doing my best over the three year time that I’ve committed to staying there. And I did stay there the full three years, I reached a point where I needed a little bit of a break. I think the summer of 2018 was the first summer that I hadn’t worked since like 1982. So I I took some well deserved time off and relax a little bit and then go right back into it when the pandemic had. So I enjoyed that definitely enjoyed my time off and I’m happy that I was able to contribute to Avaya by coming up with products like spaces, which is hopefully transforming some of their mindset and customer reach.
Dave Michels 24:30
It’s a great story because I think it’s fair to say spaces is probably the most significant development out of Avaya this century so not a bad thing to have on your on your CV there. Well, thanks so much for joining us today. It’s been exciting to learn about subspace and I think we need to talk more about network stuff as What do you think?
Evan Kirstel 24:51
Well, once you get rid of the fax machine and you really dive into packet based communication, Dave we can we can consider that but you know, you’re a bit too attached to Two parts, what we call plain old till f&e for my liking.
Dave Michels 25:03
Well, very good. All right. Well, thank you very much mo and it’s so nice to hear from you. And we’ll be watching so space very carefully.
Mo Nezarati 25:11
Thanks, Dave. Thanks, Evan. I appreciate your time today. And I’ve always enjoyed talking about subspace with you. So let me know anytime. You want some information in the conversation. Man. You gotta get out of the phone.
Transcribed by https://otter.ai