Growing a Technology Company with Oxio Co-Founder & Early Fastly Director of Product Jason Evans | 7investing
7investing

Growing a Technology Company with Oxio Co-Founder & Early Fastly Director of Product Jason Evans

Advisor: Austin Lieberman

Jason Evans has more than 15 years of experience in leadership positions at fast-growing technology companies.

Jason is the Co-Founder and Chief Business Officer at Oxio, a mobile telecom infrastructure company focused on democratizing mobile connectivity in Latin America.

From 2014 – 2018, Jason was Director of Product and Managing Direct at Fastly’s New York office where he helped the company land some big early customers in the media industry.

Jason shares how Fastly’s developer-first approach and keen focus on performance shaped their approach to developing their network and helped them land early customers against much larger and more established competitors.


Apple Podcasts Link

Spotify Link

Interview timestamps:

00:00 Introduction
01:50 What Oxio does
05:30 Oxio’s product offerings
12:55 Jason’s experience at Fastly from 2014 – 2018
16:45 Jason’s experience going door to door to businesses in New York trying to get them to use Fastly in the early days
19:00 How they began winning large enterprises with a combination of being developer-focused and using an enterprise sales team
25:00 The importance of performance metrics for CDNs from a sales perspective
27:55 What does using Fastly or a CDN do to improve the experience for users of a service like Shopify, Spotify, Pinterest, etc
32:50 How someone figured out Amazon is a Fastly customer
33:50 Can Amazon or another cloud provider just crush Fastly or other CDNs
37:00 Future use-cases for edge computing

Publicly-traded companies mentioned in this interview include Akamai (AKAM), Amazon (AMZN), Fastly (FSLY), Pinterest (PINS), Shopify (SHOP), and Spotify (SPOT). 7investing’s advisors and/or guests may have positions in the companies that are mentioned.

This interview was originally recorded on September 4, 2020 and was first published on September 10, 2020.

Transcript

Austin Lieberman

Hey, everyone, welcome to the 7investing.com podcast. Our mission here at 7investing is to empower you to invest in your future. We do that by providing a ton of free educational content like this podcast and by offering a monthly subscription service where our team of advisors provides our seven best ideas in the stock market each month for just $17. I’m excited today for our guest. We’ve got Jason Evans here, who is the Co-founder and Chief Business Officer at Oxio, so you’re doing that today from 2014 to 2018. You were the Director of Product and also managing director at Fastly’s New York office. And I think you did a few other things there as well because that’s what happens at early stage businesses. Prior to that Jason Co-founded StackPop which is an Infrastructure focused startup in 2011. Before that you, you spent 13 years building scaling and managing infrastructure teams at companies like Mediamath, Panther Express CDN, and GLG. So, Jason, thanks for being here. I’m excited to get to know you a little bit and just learn from your experience. First thing I’d like to ask is if there’s anything I left out in the intro, and then if you could just tell us a little bit more about Oxio and what you’re working on there.

 

Jason Evans 01:31

Sure, no, I think you covered it quite well, actually. Yeah, so we’re about a two and a half year old company headquartered in New York, launched in Mexico with a big Latin American focus. We we have a mobile platform that essentially allows anyone any brand enterprise developer to spin up their own private LTE instance. The idea being that these companies can benefit from owning the entire stack. They get insights into the data that can help them make better business decisions. They can control the quality of service, they can control, what sort of service and offerings they give to their customers. Mobile data, so megabytes and gigabytes in Mexico and Latin America are extremely expensive compared to what the average salary is for for people there. One of the problems is you have all these efforts and initiatives from financial products for from trying to get people banks from trying to get them credit. And there’s companies in Mexico that have loans and debit cards and things like that. But it doesn’t work if people are not connected. And so what you have is this population of people who wants to be online. People who want to be a part of the digital economy and the mobile economy, but they can’t because they can’t afford it. So one of our goals is to make it cheaper, make it more accessible. And one of the ways to do this is to to let brands and enterprises who who already have a business, they already have a customer base, they already have a goal, to let them come in and essentially subsidize or sponsor that service subsidizes service. And not like we see today with like, you sign on to a Wi Fi, they make you watch the video, like actually providing the entire mobile experience. And so we package that up for them. Make it super turnkey.

There’s a few options here and there but they control the policies, they control the prices, they control the speeds. And so it really allows companies that can benefit from integrating their product experience with mobility. It lets them letting them own that and provide that for for their customers and their future customers, we found has been has resonated super well with the Latin American market.

 

Austin Lieberman

That’s really interesting. And it’s something that’s easy as a person living in the United States where high speed internet and mobile data access is, is, as you mentioned, a much lower portion of what the average income is, than in other parts of the world. It’s really easy to forget that as much as we see these awesome things and mobile payments companies opening up these opportunities for businesses in other parts of the world. There’s key pieces there, like connectivity to the Internet, and even cell phone plans that that are just not as common in in other places. And so I know you have two products listed on the site. The first is kind of the enterprise Cloud and the second is just the analytics suite. What I would love to know is what was the reason for going with those two products? as kind of your first products as a company?

 

Jason Evans 05:17

Yeah, so so I will caveat and say Our website is, like I was telling you, it’s about two weeks out from the relaunch. It’s not super. It’s not really easy to tell what we’re doing from from the current site. We sort of since collapsed, our the analytics suite is now just part of the brand new, you know, platform. And so anyway, and I think like there’s different philosophies on like, what, you know, do you have different products? Do you have features? Do you have use cases and we had had this facet to right as for a long time, it’s like, what’s your product? It’s like, Oh, well our product or add notes as dictionaries, it’s logging. that there were people who said, No, your product is CDN. Those are features on top of the CDN or the edge So I actually think that I feel like that encompasses more of how we think about it. So our product is brand being a platform the the data and the analytics suite that that’s mostly meant so you know, going back to the brand example so So to give you a more concrete example, let’s say that we’re working with you know, Mexico they have the the tn does the corner stores are very big for example, right? And so these corner stores are a big part of people’s daily lives. And since the economy is you know, 85% 90% cash in Mexico, very, very heavy cash economy, which is also presents its own products opportunity, it presents the opportunities and the problems but people go into these corner stores to to pay top ups, right. So they’ll go in, they’ll give them the phone number, they’ll type in the phone number, they’ll say, How much do you want to put on and people don’t really understand the plans and what they’re getting for it, but you know, it could be Telsas, AT&T mobile, Star Telefonica. They give them cash, they get data and service on the client. So we’ve these corner stores get paid a small commission for doing that we’ve actually found good traction with those guys for offering their own product, right? So imagine, you know, 7-11 or Circle K offering where, or any corner shore where it’s their mobile product, right. So they’re making more margins, but also they now you know, if you think about the way the, our, the way that we designed our technology, the sim ship sort of becomes like this, this programmable device inside the phone, right?

So it’s got these very, you know, set carrier level carrier level privileges. It speaks to an app that’s on the phone and also speaks to our services in the cloud. And so now a corner store can go and do a partnership with Heineken or Corona or Pepsi or whoever it might be. Or, you know, it could be a gaming company. It doesn’t matter right, but Now you have this, this corner store mobile products. You know, as you interact with the corner store as you interact with brands, you start to create this this mobile mobile identity that’s privacy friendly, but this mobile identity where, you know, Heineken thinks you’re interesting, right, or Pepsi thinks you’re interesting. And they say, Hey, you know what I’m gonna pay for your unlimited mobile plan this month, if you’ll answer, you know, if you’ll take a survey every, you know, two gigabytes of data use, right? Or if you’ll watch this video, or you know, let us know, which is what’s your favorite product is take a picture and post it on social media. So you can start to essentially encourage behavior, user and customer behavior, the brands can control they can control that directly. And they get that one to one relationship with consumer which brands don’t often have, right, like, you know, you don’t buy your Pepsi directly from Pepsi, right? You buy your Pepsi from from Walmart, or you buy your Pepsi from from the gas station, right? And so it’s very valuable to the brands to get this sort of thing. Consistent connection. So if they’re sponsoring you for a month, they’re, you know, they’re your provider essentially for the month, right? We change the phone name, it’s sort of, you know, it’s all virtual. So, you know, you’re now on Pepsi mobile. By the way, none of these are clients. I’m just using an example. Right, right.

And you use Pepsi mobile. And now Pepsi gets the, you know, they pay a lot of money for for user research and user surveys, right. So it’s almost got a little Nielsen ish aspect to it as well, right, which is maybe a good correlation in the US. For these brands. They don’t necessarily know what to do with the data. So we sort of expose a base level analytics suite to them, so that they can start to make sense of it. We can hook into their CRM, we can hook into Google Analytics, we can we can look into different things for them. So so we present that just so that they have the ability to, you know, the larger companies probably have a pretty good system in place, but some of the companies that we work with Dell, and so just giving them some way to make sense of the data is You know, it’s like the, the full log full logging feature, right? Like, as far as NFS, he was the only one that I don’t know who has it now, but at the time, we were the only ones that had it in real time.

But a lot of customers didn’t know what to do with it, right? Like if you appoint a fire as a full logs, but like, you know, it’s super expensive to stored on a cloud log provider. And if you don’t really have a strategy for what to do with that, you know, it’s cool to have that data. It’s not as valuable. So so that was the analytics part of that.

 

Austin Lieberman

I’m excited to kind of just watch from afar and I’ll keep track of, obviously, your social media and your site and watch how it progresses because it’s a really interesting layer.

Like you said it benefits partners and companies, but also benefits users as well with affordable access to the internet, which as we’ve seen here, can do amazing things and improve quality life and stuff like that and probably even more so in in other countries around the world as well. That maybe still don’t have access to internet the way that we’re used to here in the United States. You mentioned a couple things there, where the experience you had at Fastly has led to the mindset that you have. But it feels like even with what you’re doing today, it’s still very user or developer focused. Like, you know, you talked about just people being able to spin up their own LTE networks and stuff like that, very similar to what people can do by going to Fastly website or even Twilio’s website or whatever, where developers are empowered to kind of build their own apps and do your own things without having to go through this long sales process. So I would love to start talking about your time at Fastly, you were there from 2014 to 2018. And I did a little spy research on you, I found your LinkedIn and you have this this description about your time at Fastly. It says I joined Fastly in 2014 because you believed in the team In the vision of the edge cloud, and that CDNs were not a solved problem. Over your four years, you wore several different hats at one of the fastest growing b2b businesses in the world overseeing the New York efforts and helping close strategic deals, leading to a global product team in the UK, San Francisco and New York City, kicking off new initiatives in the server to server header bidding space, and running a sales vertical focus on adtech and martech.

I’m a huge fan of Fastly it’s been one of my biggest personal holdings. I like what the business does and I really like what appears to be the culture there. If you look at Glassdoor and the CEO ratings, so I’d love to just hear about your experience there.

 

Jason Evans 12:43

I had known Artur for a few years, we were on this. I met him on this council that you remember Dyn DNS. They sold to Oracle. They’re kind of, you know, the authoritative DNS for a bunch of big companies. They had the big outage a few years ago. which most people know them from. But no, it was a great company anyway, they had this tech council for their customers. And so I met Artur on that. And so, you know, I was sort of fascinated by Fastly from, I guess this was like 2011. I was still at Mediamath. So like, 2011- 2012 like super, super early. And so, you know, we would chat. Then he ended up hiring, the CTO of Dyn Tom Daly, who was also he was an advisor. He was an advisor for StackPop, he was a good friend of mine. And so I was like well, this guy is smart and just like talking to Artur I mean, I don’t know if you’ve ever you probably watched his talks you know, he’s just like, he’s just sort of next level intelligent when it when it comes to understanding like, what happens in a computer, right? And that was one of the reasons that that those guys were so good in the beginning is like, they just thought about that. performance on a whole nother level. You know, I think about like the first CDN I worked out was was called Panther. Artur was a customer of Panther when he was at Wikia. And he was one you know, when you have when you’re an infrastructure provider, you have these customers who, you know, like as soon as you have a blip, or as soon as you have any sort of production outage or not even outage but no issue that they’re going to be right on top of it. Right. They’re monitoring, they’re paying attentions. And so Arthur was one of those customers. And he was always frustrated by the fact that it took minutes to purge content.

Keep in mind at the time, it took akamine hours, like if not a day to purge content entirely from the network. So this concept of of instant cache invalidation was as far as anyone was concerned that business impossible. So like, I remember a ticket from long time ago, arch and I laughed about I guess later on, but He was like, why does it take five minutes to purge my cash? every five minutes is really good. Like, it’s so much better than what’s out there. But, you know, that was essentially one of the drivers applied. He built wikia, I’m sorry, he left he built fastly, you know, so he’s on that project while he was CTO Wikia. And so just just the fact that, you know, Tyler and Simon and Jason and Artur and his guys did that was just fascinating. And so like, from a technology standpoint, like, I feel like fassi always started like something just on a completely different level than anyone else. And, and so and, you know, I liked Arthur’s personality. Obviously, Tom being there was great, and he was sort of brilliant infrastructure and networking mind. And so I kind of always I had my own company at the time. And so it just wasn’t time. But yeah, I always sort of wanted to join if something happened and then the the opportunity came up after we started transitioning out of stock pop. And I just kind of reached out and said, Hey, I want to join the company. I don’t really care what I do. They asked me if I wanted to live in San Francisco or New York and I said New York’s biggest better for now. And so that’s so I became a managing director of New York which was sort of a hybrid sales tech overlay you know, office, you know, sort of region management sort of thing. But yeah, so I spent most of my time though, that first year on the on the commercial side and the customer side, so you know, literally just going door to door in New York. Media was great for us because…if you think about media, right, like when you poll if you’re Vox or BuzzFeed or, you know, Business Insider, whatever. If you can only cache the, if you can only cache the images, like it just kind of sucks, right? Like you’re always blocking yourself at the root domain. If you can’t cache cnn.com or if you can’t cache buzzfeed.com, then you know, it’s just you’re just going to have an experience where like, okay, yeah, your images are at the edge. But like all this, you know, all this HTML. In the roots, the root asset is going to take forever to load. So because of the the instant cache invalidation, right, so the sort of base number is 150 milliseconds. Your your content is purchased globally, right. And so and so no one else could do that. And so what fascinated was, they had the instance, and validation. And then they also built with the developer mindset, where you could use API’s like the the purge API was just an HTTP endpoint. So you just go into your CRM, and you publish a story and again, the danger was okay, if you did SAS or HTML on on akamine at the time, I’m not gonna pick on akamine but this is a common use case, you you could it was too dangerous because if you made a mistake, or if you had to issue a correction, it would literally take hours around the globe to, to, for that to invalidate, right. But with fastly, since it was instant, you could put out whatever you wanted text, whatever. And if you needed to make a correction in your what we would tell people is in their CRM, whatever, they’re not CRM, their their content management. So CMS CMS, in your CMS, just call the call the purge API, whenever you make a change, and you know, we’ll go faster, we would purge the cache and it would be instantaneously populated so. So media was like a field for us. So once we got a couple of good customers. I think I like Business Insider was definitely one of one of our first and it was, you know, at the time of a really amazing deal. And so, you know, we would just piggyback off off each of those.

 

Austin Lieberman

What were those early conversations like as a kind of scrappy company without a proven product yet and what did it take to convince them? Did they have to see it? did was it just really just like, give us a shot and try us?

Jason Evans 19:22

That’s a good question. You mentioned this earlier in about the philosophy at Twilio and Fastly, and companies like that, like, I’m a huge believer in like, you need to have a, if you’re going to be an enterprise software business, you need to have a product that developers love, and will will use no matter what and then they’ll evangelize that up, right, they’ll push that up to the organization. You also need to have an enterprise sales team or a sales team that’s also working that from from the top down right in that and to me that that formula is is the only form In 2020, of how to build a software business, and so Fastly had done that, right, like Artur and those guys had been super involved with, with, like, the the tech conferences for a long time. So they had a good core base, right. Like, I mean, like, you know, all the, you know, Simon was really good friends with the early Slack, guys.

Right. So Slack was an early customer and so we had enough validation from customers that were respected from a technology standpoint and then we had the product which developers could use and like. At the time, at least, I don’t know now, but there was no way that you were going to get an Akamai account without spending hours on the phone and, you know, hearing sales pitches and things like that. And so, yeah, Joshua Bixby has been, he talks about that still on, even on some of the earnings conference calls and different things. COVID has actually caused, like the “steak dinner” sales to go away. And now it’s truly becoming like the best of the best because you might meet over a Zoom and you don’t have to kind of wine and dine these different customers. So they’re still talking about that today.

 

Jason Evans 21:21

Yeah, it’s totally true. I mean, honestly, like, back to your earlier question, like, with COVID, like we were ocsea was was much more efficient in terms of like, you know, because when you go down to Mexico City, like Mexico City’s a, it’s a great city. It’s hell on traffic. So if you have a meeting in Santa Fe to Polanco, like, you know, good luck getting there in less than an hour and a half and, and then you know, the culture Mexico’s like three hour lunches with lots of tequila and things like that.

 

Austin Lieberman

I spent time in in Jordan, and it was the same the same way very long meetings and food and stuff like that. It’s fun, but it’s Yeah, it’s not the most efficient.

Jason Evans 21:58

It’s not the most efficient so so Yeah, to answer your question we we had we had good validation from some some early power customers like Twitter was, for a long time our biggest customer and and I know they had been working on that account for a long time like slack was there at the end. Yeah, like the Etsy guys were kind of like, speaking of Signal Science like that the fancy guys I think came out of the signal science guys came out of exit early on. We had a product that developers lights that were interested in, so we would go in through them most of the time. And yeah, just as we It wasn’t easy like it. Yeah, it was pretty challenging at the time. I mean, we didn’t have the network capacity.

We didn’t have security products at the time. There was you know, Akamai was still very, very present in almost every deal. After my they started playing ball earlier than I fought on the price side of things. They you know, they had a they Because to name their price for a long time, especially with media companies who weren’t, you know, videos a little different because there’s so much bandwidth, but the media companies like people are overpaying and but it wasn’t easy. But we had the we we made the developers that the early team built a product that developers love to use. And so yeah, you go in and you’re like, Okay, put, you know, put this up on your put this up on the website. Okay, hit purge, refresh, boom, it’s gone. So, you know, it was an easy demo to show them that on the media side.

 

Austin Lieberman

Yeah, Joshua Bixby has talked about this a little bit as, as, especially now, we’re seeing Shopify as a customer and how fast they’ve grown and even he didn’t say this directly about TikTok, but one of my takeaways about TikTok, you know, I think this is seen as a risk to Fastly because I think it’s 12% of revenue for the last six months came from TikTok, half of that was from the US revenue. So, you know, after the last earnings, there’s this big fear of losing that revenue, but I’m looking at it and I’m like, this is the best advertising in the world because this is the the one of the fastest growing probably the fastest growing app in the world, and it’s they use Fastly so you know, whether TikTok stays or goes, why wouldn’t the next app that wants to grow exponentially use Fastly after they know that TikTok was able to do the same thing. I’m sure it’s more complicated than that. But that was kind of my high level takeaway. So it still seems like Fastly is proving to everybody that that, you know, performance is is top notch.

 

Jason Evans 24:53

Yeah, another thing that we were speaking of which, like if I put on my like sales hat from from a long time ago, We would constantly and I guess, maybe I don’t know if you looked at the the blog stuff, I think I wrote a couple things about performance, we would talk about that a lot. We would we, we really tried to create the story that performance is not just how stat how the milliseconds it takes for a static object to respond to a third party monitoring service. Right. That’s how, you know people used to use I mean, they still do with companies like Catchpoint is great company. Gomez was one back in the day, but you know, basically what you do is you would say, Okay, I have three CDNs. I’m going to monitor this object from each from multiple different cities and multiple different carriers so I can understand what the response time is from from different regions and over different carriers. And um, you know, clearly that was easily gameable, right. So if it’s just a static object sitting in RAM, it’s not that interesting. But you know, so when we were talking about performance, we would say performance is also like, how How long it takes you to invalidate your cache and purge your cache? How long does it take you to push out a config change? Right? That was another place where we had a huge advantage over Akamai. Every user had full control of their VCL configuration, which already on its own was super powerful and super granular. And so you could push, you know, you could literally push out your configuration. It wasn’t milliseconds, it was seconds, right. So you screw up your configuration and things start failing, you could roll back the latest version, like it literally in literally in seconds. And so, you know, that’s, that’s performance, right?

Like performance is also being able to see what’s going on at the edge, right. And we were the only company the only CDN edge cloud. We didn’t call ourselves edge cloud that can By the way, what we were the only CDN that could give you literal real time logging. I use this story at Oxio, but instead of your CDN being a black box we could give you true visibility on what was going on at the edge of your network with the logs. And so we kind of, you know, we started to create that story of performance is all of these things, not just a third party response time that that’s easily gamed level.

 

Austin Lieberman

We’ve talked a lot about the approach or the understanding of Fastly from a b2b perspective, or what Fastly’s direct customers might experience. But for people that are maybe just everyday Shopify users, or Spotify users or any different customer of Fastly, how does Fastly’s CDN impact their experience as a user?

 

Jason Evans 27:43

Tradsitionally, CDNs were really just for static objects, right? If you look at sort of the origins of some of those companies in the late 90s, early 2000s. So you know, essentially, what a CDN does is it moves the content closer to the end user And it, it makes the entire experience better, right? Because I mean, no matter what you do on the internet, right, you always have this the speed of light problem, meaning that if you’re 5000 miles away from a piece of content, like, you look at the speed of light over fiber plus a little network overhead plus some fiber overhead inconsistencies, and retransmits, and things like that.

And, you know, it just it’s it’s a poor experience, if you have to, you know, if you have hundreds of milliseconds latency every time you need to do something, right. And so the traditional role of a CDN was just to to get content closer to the end user for the reasons we talked about earlier. It was typically only static content, right? So yeah, so when you think about a Spotify and a Shopify and some of the sort of power customers that Fastly has, and they had this goes for, I’m still blown away. I mean, congratulations to the Shopify guys. I’m still blown away at how fast that company grew and how it grew. But, yeah, I think one of the things that describe those guys is I don’t I don’t know like I I know some of the Shopify guys, but sorry, Spotify guys, but what Fastly did like, again, like with, with extending code with extending code to the edge right, in the form of VCL, the configured varnish configuration language, what they did with making everything instantaneous, is you now like, it felt like your own infrastructure, right? And so there was this sort of visceral feeling that you control. When you when you have access to like this huge powerful network, it can protect you from DDoS attacks can do everything fast. And you know, when it’s super responsive and super configurable it, it’s a huge advantage. And so I think for guys, like, you know, if you think about Shopify, like, clearly like in order for them to be able to cache things, and they have so much dynamic content and content that sync and you know, that’s not consistent from one user to the next, and prices change and things like that, like it’s a great use case for profesi because it can instantaneously get rid of anything that you don’t like, right? It can serve up a different object based on the request itself. like where’s the user coming from? What is what’s in their cookie? What’s their language preference, things like that. So, so yeah, I think that that’s what, like, when you when you spin up, Spotify, I don’t know. Exactly how they how they’re using fastly today. That and but like, you just you just don’t have access to. And again, I haven’t followed this as closely in the past year. So I don’t know exactly what the other edge cloud companies are doing. But you know, you just didn’t have those, those options with the other guys. And so it was more of a, it was sort of a legacy like, and I know people like insult older CDNs by using that word. But, you know, there is like, there is a new way of publishing content on the web, and some companies are there and some companies are not, but but I think, you know, giving developers the power to, to make that feel like your infrastructure was was something that really worked.

Austin Lieberman

My takeaway is that it feels like as a user out here, all these customized playlists that I get from Spotify, or if I’m shopping on Shopify, prices that are updated and accurate inventory that’s updated and accurate, and basically customized down to where I’m at my preferences and things like that. That’s how I’m able to get those types of what feel like unique experiences from all these different services that I use are from basically CDN and edge cloud. Like that’s the improved user experience, I think that comes and then fastly, I believe, is one of the best at providing us. So I just wanted to hear from your perspective, from somebody that’s actually been there and done it versus me just kind of reading about it.

Jason Evans 32:29

So the cool thing about CDNs do also like you can, you know, if you open up your dev tools, or if you, you know, open up a Wireshark or sniff your traffic, you can actually see a lot, right so like, you know, we’re like, the person that exposed that Amazon was customer and stuff like that, like, it’s all just kind of public information. So if you’re ever curious, and you have Dan, he’s super good guys. He’s been good faster, but you know, if you’re ever curious of what’s going on, you can Put, you can take your your Spotify app and put it through like a local proxy on your laptop and then just sort of, I mean TLS makes it a little harder if it’s encrypting everything, but you can kind of get an idea for what’s going on under the hood. And like how, how that particular customer application is using its its CDN partner. Yeah, yeah, that

 

Austin Lieberman

That was really interesting to see that Amazon was using was using Fastly, but Fastly has not talked about Amazon as being a customer yet. So it’s obviously not at that point where Amazon has given the approval for that. If it’s official or not, which neither of us here are saying that that Fastly has said that. All right, so why wouldn’t an Amazon or a cloud provider just just be able to crush Fastly or any CDN?

 

Jason Evans 33:49

Yeah, we got this question a lot early on, and like, I think, I don’t know, I think a lot of a lot of investors ask this question. I think it’s a little lazy. You know, it’s just hard for somebody to be good at everything, right? AWS is an amazing platform and product. It also does everything right. And so, you know, so there’s a lot of like, there’s definitely some, you know, like, s3, in my opinion is a much better product and an EOB your a lb load analysis stuff. So and you know, I think part of his focus, I do think that the, you, you just you don’t get the, the brain and the technical, the technical advantage that, that some of those guys that are turned and a lot of those guys gave fastly in the beginning and that continued on right when you start out 10 steps ahead of anybody before before they move like it’s easier to stay a little bit ahead. So, you know, I think part of it was focus and you know, when you don’t have a you know, thousand purpose network, right. It’s It is easier to do 123 things well, right. And so that just sort of obsessiveness on performance. And that’s one of the things we’re like, with from it just CloudFlare business model. And by the way, I, I do have my beefs with them, but I think cover does do some things, well, they do release, they release a lot of products, they’re really good at releasing new products. They, you know, as far as I’m concerned, like, they’re the closest thing to fastly in terms of edge logic in putting, you know, pushing out, pushing out code that can be run on the edge servers.

But it’s also like, you know, when you have 10s of thousands, I don’t know how many customers they have a hundreds of thousands of customers or something. You know, like, there there is like, there’s physics there. Right? And so, you know, if you have hundreds of thousands of customers, that’s no, that’s, that’s heavy from a configuration standpoint. If it hurts your cache, the cache ability, right, one of the things There’s a guy mazing do fancy code, who actually, who does performance stuff for them? It’s really a real HTTP wizard.

We would monitor how long an object a non hot object would stay in a cache. Right? And that’s a very important thing humans that hasn’t talked on that actually the republic but when you look at you know, when something when when objects falling out of cash in five minutes like it’s that’s a real problem, right, like, sit when it got below six hours, I would get worried. And so I think that I think that there are there are trade offs in approaches on that.

 

Austin Lieberman

As far as expanding use cases for edge computing, where do you see that going? In the next few years or five years, there’s been talk of autonomous vehicles and things like that.

 

Jason Evans 37:41

Yeah, that autonomous vehicles are always kind of the the, the use case people use. Yeah, I think that there’s there’s a really good talk. I’m pretty sure it’s on on YouTube. Tyler momoland is CTO of Fastly another another brilliant dude. And he He gave a talk on sort of the future of edge computing. And, you know, he brought up the obvious, which is like, like, edge computing is here, like, it’s going to get better. There’s, there’s, you know, the IoT use cases, the car use cases. But you still have to deal with the data, right? And so, you know, it’s you can, you can do logic at the edge. But if that logic requires querying a large data set or having large data sets to choose from at the edge, then you know, you’re going to be super limited in that and, and I think that’s, to me, that’s one of the big unsolved problems. And one of the things that is holding back the use cases is, you know, think about how much data we’re generating now how much is collected, how much like the average enterprise or startup stores and gigabytes, terabytes a day. And so that the data storage issue is still one that I feel like hasn’t, it hasn’t been solved to a point where You know, it’s, it’s holding back the the sort of explosion of edge compute. So, you know, I think simple simple decisions can be can be made at the edge, right? Like if you know the the response, if you just need to let someone know, okay, I heard you. By the way, I’m going to take this data that you sent me and I’m going to asynchronously load this into a data warehouse for you. You know, there’s a lot of use cases like that. I do think with this kind of the stuff that oxy is doing on the private LTE side, you know, as metal and hardware gets cheaper and cheaper, as it always does, you know, you can start to put enough out there, right, you can start to sort of figure out which data sets should be loaded into, into which regions in which data centers and things like that. But, yeah, so so I don’t have a magic answer of what’s the next big thing for edge computing. But I do think that we’re still a ways off from seeing a huge, huge like groundbreaking change on it.

I think mobile will drive it. And as the as the mobile ecosystem changes, then you know who who’s operating that I work at the edge and who’s operating it at the core. I think that’s gonna that’s going to shift and there’s a lot of opportunity there for companies to, to be in that path so that it’s not just this big black box network, big black box CDN, and then your, your AWS instance.

Austin Lieberman

Well, Jason, thanks for your time. Everyone that’s listening. You can follow Jason @jasonhevans on Twitter. We’re here to empower you to invest in your future. Thanks for listening.