Q&A: Zoom CTO Brendan Ittelson
2020 was a rough year for just about everybody. Political chaos. Wildfires and floods. Job losses and economic stress. And, of course, a global pandemic. It feels insensitive and downright crass to try to find any bright spots in all of that darkness, but those of us in the streaming video industry can take pride in the fact that we made all of that a little more bearable. Can’t go to the movie theater? There’s no shortage of blockbuster films and binge-worthy television available via OTT. Can’t travel for business? Videoconferencing is the next best thing. Can’t go to school? Kids and adult students across the country have become intimately familiar with video classes. And while there’s no replacement for hugging your family at the holidays or getting together with your friends for a few pints, all manner of apps and services are available to let you at least see each other.
And for now, at least, one service rules them all, to the point where its name has become synonymous with video meetings: Zoom. Zoom has become the default option for business meetings (some would say it’s the default to a fault—not every call needs to be a Zoom call), virtual conferences, and video happy hours. Zoom is easy to use, offers a free tier providing for calls up to 40 minutes as well as inexpensive plans with no calling limits, and has robust webinar capabilities. And—most importantly for readers of this magazine—the video and audio are rock-solid compared to most of its competitors.
The numbers are staggering. In December 2019, the company saw 10 million daily meeting participants. Since April 2020, Zoom has reached, on average, 300 million daily meeting participants, both free and paid, and it ended the third quarter with an annualized run rate of 3.5 trillion meeting minutes, according to a company spokesperson. During the company’s Q3 2020 earnings call, CFO Kelly Steckelberg said the company had a 367% year-on-year growth in revenue and a $3 billion annual run rate. Subscriptions to new customers accounted for 81% of that increase. Some of those are personal subscriptions,but Zoom ended Q3 2020 with more than 430,000 enterprise customers with more than 10 employees, a 485% increase over Q3 2019 and accounting for 62% of total revenue.
Zoom’s growth hasn’t been without challenges. Reports of “Zoombombing,” leaked passwords, and other security concerns arose in April 2020, though most of the issues turned out to be due to users’ incorrect use of settings. The company has made significant strides in addressing other privacy and security concerns (including a settlement with the FTC regarding its use of the phrase “end-to-end encryption”).
I sat down—via Zoom, of course—with Zoom CTO Brendan Ittelson to talk about the company’s remarkable growth, its plans for the future, and a little bit about its "secret sauce" for great video and plans for live streaming producers.
Streaming Media: Why do you think Zoom took off the way it did in the wake of the pandemic, as opposed to other solutions on the market?
Brendan Ittelson: I think it is a combination of strong execution in terms of a product and really being customer-centric in our design, and you know that has called out to individuals. We are so user-experience-focused that I think people have found value in that ease and simplicity.
On the technical side, there are a lot of things—our cloud architecture, our geographically distributed ability to scale, and how all those play together—but I think at heart it really comes down to that strong user experience, because you know, given the world that we’re in right now, the ability to connect is so critical. Being able to have a tool where the technology disappears and we can focus on communicating—that’s key.
What is it about the technology that you’re using in your approach that allows the technology to disappear?
I think a lot of that is being user-focused in our design. We have a very engineering-focused culture, but I like to refer to our team as compassionate engineers. They’re always doing problem root cause [analysis] before they go to solution. And then when we think about those solutions, it’s looking at the human element and how to really focus on that individual. Then we dive in behind the scenes to build that technology to make that work.
In scalability, where we’ve got our distributed data center network and our multimedia routing platform that really allow the optimization of connections. When you combine that with our multi-bitrate encoding, that allows us to adjust the quality of service for every participant based off of a number of factors—each person’s experience is customized for the environment that they are in. This helps us deliver a reliable, customized experience to every meeting participant.
Can you talk a little more about how that cloud architecture and your networking work together?
From a high-level architecture [standpoint], we leverage 19 data centers around the globe, as well as cloud data center providers, to build our global network. It is truly a global network when you look at all of our different points of presence so that an individual can connect into Zoom’s back end anywhere around the world. We can then optimize their connection across our network so that they’re getting the voice, video, and audio that is catered to their system, their CPU utilization, packet loss, and jitters. So, we really optimize to the edge with that conductivity.
Are you using outside CDN or cloud partners? How much of this is Zoom doing in its own data centers, in its own networks, and how much is out there on public networks?
It really depends on how our customers are leveraging our product. Take our global network of data centers. When you’re in a meeting, you are connected to that global network and using our meeting servers. However, we do have the ability to do live streaming out to any platform that supports RTMP [Real-Time Messaging Protocol], so really, it’s a question of how folks want to consume the media and have that interactivity, or that broadcast ability. If they want to use a CDN or use our network, we really believe in offering flexibility there.
You talked earlier about scalability. Even though you were well-positioned before the pandemic hit because of what you call the “compassionate engineering” approach that you take, you couldn’t have anticipated this kind of scale in 2020. What did you have to do to meet the demands of so many users?
We’ve always focused on simplicity and are designed to allow us to grow very quickly and ramp up. Our architecture was built so that we could add additional capacity in our data centers and public cloud providers quickly and efficiently. So our team was able to constantly look at demand day to day and grow the resources across that global network to help keep up with the load as it came in.
Let’s talk about the future. As people start returning to working in offices and traveling for business, do you think that Zoom usage is going to go down significantly? Or do you think it’s become such an integral part of how we work and live that we’re going to maintain this high usage?
As we look toward the future, we are starting to define what a new normal is, and I think that is still some ways off. But I think a lot of folks have taken away preferences from this experience that are going to continue and evolve as we go back to work, such as the idea of hybrid sessions [meetings or events that combine in-person and videoconferencing]. I’m excited to see the continued innovation that’s going to occur here.
Zoom was criticized in early 2020 over security and privacy concerns. What has happened in the last 8 months or so to make the platform more secure, either from a technology standpoint or from a user education standpoint?
We take the privacy and security of the platform extremely seriously. On July 1, we completed our 90-day security plan where we focused all our engineering effort on continuing to up-level our product in the security space. During that time, we froze all feature development and focused on continued improvements in our platform—things like Zoom 5.0 with the new security icon, the defaulting of settings such as the waiting room and passcodes, our acquisition of Keybase and rolling out our end-to-end encryption solution, as well as the data center routing configuration to actually let people choose how they want to route through our global network.
You put all that together from a technical standpoint, as well as forming our CISO [chief information security officer] advisory council and continuing to up-level our bug bounty, and I think it’s very exciting. We’ve been able to make progress to continue to enhance our policies, our procedures, and the product, because at the end of the day, we want to safeguard our users and their information.
Your competitors aren’t sitting still. Google Meet has come on strong, and Microsoft Teams has made enhancements. How is Zoom setting itself up to continue to be the leader in the space and compete with those changes from a technology standpoint?
Innovation is one of our key areas, and really it comes down to our user-centric approach and being focused on them. We’ve made investments in data centers [in 2020], we’ve brought on new [personnel], we opened new technology centers in Pittsburgh, Phoenix, and Bangalore. A single-minded focus on providing the best video communications experience to our users is key to our DNA, as well as constantly listening to our users and understanding them. You’ll continue to see us executing in the way that we have.
Streaming Media readers are always interested in encoding and codecs. What is your primary codec at this point, and can you talk about the value propositions of the leading codecs on the market?
Our primary codec is actually something that we developed internally for Zoom. That ability to do the multi-variable bitrate and achieve the quality that folks see within a Zoom meeting is based off of that codec, which we’ve spent many years to develop and perfect. But we also believe in being a very open platform, so we have the ability to interface with different systems and different devices and transcode as necessary between those different codecs when working with different services.
I can’t go into a lot of details [about the basis for our codec]. But in the design, we looked at what’s out there, what’s the best technology, and how can we improve upon it and take the best elements as well as innovations of our own to further raise the bar.
Is Zoom going to become more live event producer-friendly in terms of things like SDI input and output, as well as NDI?
It is definitely something that we’re looking at. Going back to that customer-centric approach, we have gotten the feedback [that producers are interested in that]. It’s something that our engineers are looking at. We want to make sure that we are nailing the use case and really making sure that we’re not just delivering an NDI feature, but delivering something that has the right context and is a full package for those individuals for their workflow to make sure that they have that amazing experience where it just works exactly the way that they’d expect it to.
Are there any other technology improvements on your road map that our readers should watch out for?
We’re growing in a lot of different areas. At Zoomtopia, we announced our “On Zoom” platform, where folks are able to have events in a public directory and have the ability to monetize them, as well as developments in the API and SKD space, where folks can embed the Zoom technology within their own applications or do a deeper integration between their platform or product and Zoom to give that great experience and the best of both worlds.