JFrog Ltd. (FROG) Earnings Call Transcript & Summary

Good morning, good afternoon, good evening, everyone. Thank you for joining us today. We are here for managing IoT software at Scale. I have the great pleasure in introducing Amit Ezer and Yoav Landman. Amit is CEO of Upswift, and we have Yoav, CTO of JFrog. So I just wanted to give a quick introduction to the both of them. And versus me chatting away, I'm going to hand it over to the both of them. Thank you guys so much for joining us today.

Okay. So welcome, everyone. Let me just give you a brief overview of our agenda for today. We'll give a brief introduction about JFrog. And then I'd like to speak with you about the challenges with updating connected devices. Next, we're going to move on to Amit, who will speak about the benefits of Upswift and how they solve those challenges for the world. Of course, we're going to show you a demo and hope is that we'll walk out smoothly. And finally, we are going to speak a little bit about what's coming up next, what JFrog is going to do with Upswift. We're going to take this domain of connected device updates in JFrog. So let's just dive right into it. So JFrog is solving the problem of what of a vision, we call, Liquid Software. The idea is to have continuously updated, what we call, versionless software. Of course, every software has a version. But in the eyes of the user, the update needs to be so smooth and transparent that the version doesn't matter. And you can think about it as software flowing seamlessly from source to destination through the run time. And since we speak about the run time, I'd like to speak a little bit about the challenges in updated device -- in updating devices. So a couple of challenges here. I'll start with the fact that the act of deploying a release to an edge is often a siloed process. So it's siloed from the traditional CI/CD process. And it's often looked at as a separate state, different tooling to achieve this state. JFrog customers actually approached us and asked us to allow them to continue this flow of the DevOps pipeline all the way to the connected devices to the world of IoT. So this is the first challenge. Another challenge is about securing the update. So you need to have to apply the security from 2 directions, right, from left to the right. First, you want to make sure that the software that you're running is secure, you want to scan it, you want to rule out any vulnerabilities or compliance issues. And then once your device is already running the update, you want to validate that it behaves the same way that you expect it to behave. And one key differentiator from traditional software is that if you have a faulty release, it may break the device. And then fixing this is a much more acute problem than in traditional software because your device may not be so easily accessible. For instance, it may sit at the top of a communication tower or something like that. And this leads us to the -- security leads us also to the next challenge, which is maintaining a chain of custody for your releases. So a strong side effect of being a disconnected flow is that you also lose and you get disconnected from the software supply chain process. And then it means that the binaries and the releases that end up running on the device cannot be easily tracked to the process that's created them. They cannot be vetted against the software supply chain, and this is a [ miss ]. Nondata center. Nondata center is something many people kind of look over it, but updating devices is really, really different from updating software that is running on a data center on an immutable infrastructure data center. First of all, devices, normally they cannot be connected -- contacted directly. They don't have an open port they listen on. They may be disconnected from time to time. More often than not, they are behind some kind of a double net. And they -- I call this a cattle of pets, right? Because you treat your devices at least the way that the scale of the devices, the number of your devices is like a cattle. But each device is identified. So it's like a patent. It's not -- it's a non-replaceable piece of run time. You cannot just fire up a replacement device that you do with the infrastructure [ code ]. It doesn't work for devices. So that's another challenge. Another challenge is the scale, right? So if you are a company, even a small, medium company getting success in the world of IoT, you will find yourself managing sometimes even in your early stages hundreds or even thousands of run time of devices, which you would otherwise not encounter as the DevOps of a traditional microservice-based deployment. And finally, there is the question of what's happening post update. So you can get your software on the device. But like similar to any kind of software, it doesn't end up there. You need to be able to monitor the health of the device. You need to be able to roll back a bad release, if needed. You need to be able to run remote troubleshooting, get the logs and extract them, even remote controlling to the device to see what's going on. So this continuous feedback loop is even more challenges -- challenging in the scale of IoT. And what we are saying is that -- and we're already seeing this exceeding trend and this fast-growing domain in the market, that companies are slowly becoming edge companies. Almost every company will end up using an edge technology or employing edge technology in order to compete in the market and it's kind of a new necessity for success. It's happening very, very fast. And we broadly speak about 3 types of edges of -- 3 types of connected environment. First is this new consumer edge. It's used to deliver software closer to users. Think about users playing a game, basically letting the users interact with your software faster. Then there is a very growing area of regional edges. This is where you deploy your software into regional locations, such as local retail branches. And this has recently been pretty much heavily accelerated by COVID, and we don't think this trend is going anywhere. So COVID accelerated a new type of application-driven businesses. For example, this whole curbside pickup that you go to the store and people deliver the -- via the application the merchandise to you. And we think we're going to see more and more of that application-driven businesses. And finally, there is the third kind of edge, which is about connected devices, from routers to wearables, think about Nike, moving from an apparel company to a smart fitness company. So this is the way that you actually use connected devices to extend your TAM. And many of these devices run embedded Linux, which is what Upswift is predominantly dealing with. And we see tremendous innovation and use case variety happening across the Upswift customer base. So this is the third kind of edges. This is -- what we're seeing here is the JFrog DevOps platform. It's one of the most widely adopted DevOps platform covering all the stages of software releases -- software binary release management. Normally, you start on the left side, you have some kind of a source control system like GitHub. And in most environments, the first thing you do, first, you consume some binaries dependencies that you assemble on your application. And then you compile and package your software and bundle it into pieces of binaries that are going to go through the different stages. So testing, release approvers and distribution and deployment to the run time. The end state is that those binaries are going to run and end up in your run time. And the JFrog platform allows the development -- allows developers to manage binaries across all those stages. And more specifically, we have Artifactory that stores and manages the life cycle of binary Artifacts like software packages and container images. We have x-ray that scans the binaries to clear out any security and compliance issues. And we have distribution, which is about taking the binary and moving it across the physical boundaries closer to the run time, including to devices and enabling deployment. So in summary, what JFrog offers is an end-to-end DevOps platform for managing the life cycle of binary releases from code to deployment to production. And the issue with Upswift is that it's actually a very natural extension to the DevOps platform that JFrog provides today. So while for devices today, the DevOps flow, not the JFrog, it stops at the distribution, and then there is a new siloed process that is starting. This natural extension of the JFrog platform by Upswift allows the continuous flow to be carried on to update, to run, to monitor and back again. So it's a continuous loop. And now I think it will be a good time to pass it over to Amit. I think you're all waiting to actually see the product and see Upswift in action. So to you, Amit.

All right. Thank you. I'm very excited to join this webinar today. Actually, this is the first time we share some insights about what we have done at Upswift and how our mutual future with JFrog will look like. So thank you all for joining us today. I hope you will find this webinar and platform demo that I'm going to show in a few moments valuable and informative to your future projects. And feel free to reach out after this session, if you have any questions. I will leave my contact information in the chat to you all. So I'm happy to begin with a quick intro of Upswift product, vision and market. Upswift's goal is to bring every edge device from robots to connected cars, security cameras or kiosk screens to the cloud. Bringing the edge device to the cloud means to give IoT producers, the developers, the power of managing the device as if it's in their own office, in their own local network. Today, there are already tens of thousands of devices all over the world managed by Upswift platform. And it's just amazing to see so many different verticals and industries that are doing this transformation for building connected devices that are being managed remotely. We see health care machines, agricultural devices, automotive products, and the list just goes on. Developers today prefer to manage the fleet remotely and ship software quickly and safely in each and every industry or edge device. Great. So let's for a minute talk about Upswift's purpose. Our solution solves the complexity of managing devices with the plug-and-play design that can be integrated into any type of Linux-based device in seconds. And it brings 4 major values to the developer. First and most importantly, Upswift helps to reduce or even eliminate product recalls. Take, for example, thousands of small kiosk screens at post offices across the U.S. that are suddenly required for a recall because of a software bug that occurred. With Upswift, these devices could be completely supervised remotely. Even if the application crashes, the developer can always have access to support and fix the issue remotely. The second point is ensure software reliability. If it's a fleet of cars or traffic lights, the device's software must stay stable and reliable 24/7, and that can be achieved by pushing the daily software updates to the edge devices. And when the device's fleet is stable, we can start thinking about remaining innovated. So just like in today's cloud deployment flow, with Upswift, your edge devices will stay updated with the latest feature releases the development team worked on, simple as that. And finally, when the product fleet is in the field, in production, it is important to always stay in control. Here at Upswift, Upswift is continuously monitoring the device resources and application, helping you, the developer, to analyze, find device software issues, debug and fix everything before any real damage occurs. From here, let's jump to the demo of Upswift platform. So I will share my screen. All right. Great. All right. Before we jump into the major features of the platform, we can notice from the categories in the left side, right here, that the platform includes the necessary tools to maintain a fleet of devices remotely, starting from over the air software updates, capabilities to control, monitor and security, and it's all mixed with smart alerting system. And it's of course all designed to work out of the box, just [indiscernible] agent on the edge device. It is important to notice that the agent is always running as a client side in terms of networking, which, of course, is great for devices that do not have public IP and running under double net. For this webinar session, I will use a Raspberry Pi device with a touchscreen on it, running a blue demo application called webinar V1. Let's go back for a moment to my camera. All right. So right now, if you will look closely on the camera, you will see my Raspberry Pi device with the application Webinar V1. And let's go back to my Upswift account, continuing the demo. All right. So back to my screen. All right. That's it. We are ready to begin. Let's begin with going to our group for today, our specific Raspberry Pi device. So the device will be under a specific group called [indiscernible] right here. And let's jump to the device stage, where we will see our device today, right? Here it is. So here is the main page of Upswift platform, where we can simply view the devices in different groups with name tag and most importantly, live status, if the device is currently online or offline. So as you can see right now, my Raspberry Pi is online, and we are going to use it right now in the demo. So I will focus on 4 of the platform tools, updates, remote control, remote commands and some of the monitoring capabilities we have here at the platform. Let's begin with the updates category. All right. So Upswift update is a lightweight, easy-to-use tool to deploy software to the edge devices. It's all based on the creation of a flow, right here. In the flow, you can build the steps of the deployment process. For example, you can run a bash command, deploy a file, defining the rollback, integrate a Git repository or even a docker registry. It's all very intuitive and easy to use right from here. So for this demo, I have created 2 different update flows, both pulling different images of containers from JFrog Artifactory and pushing them to the edge device, while replacing the current running container with almost 0 downtime. So for a moment, let's jump to the Artifactory right here, where we can see both of my images, V1 and V2. So V1 is right now running on the device. We have a running container on the device running webinar V1 application. And our goal for today is to deploy the container -- the new container V2 to the edge device to my Raspberry Pi. So let's go back to my -- to the updates category and begin the deployment. So to begin the deployment, all we need to do is just to click on the new deployment button right here under the role of webinar app V2, choose a group of devices or a specific device. In that case, we will use the filter right here to choose a specific device and then we're going to go ahead and choose the Raspberry Pi and click on apply. Of course, we can deploy updates on a group of devices at scale. Let's click on next. Of course, the total affected devices will be one, and we'll click on finish, choose our webinar app for this demo and the new version that we are going to deploy. And then let's click on create deployment. And go to the progress. So right now we are waiting the device to receive the update in a few seconds. All right. We're going to see it in progress. Let's quickly move to my camera again, just to show you how it looks like. It's pretty amazing, actually. See. So you can see right now that my Raspberry Pi screen changed to the new green application, webinar V2, pretty amazing. Going back to my screen. All right. We can see the output of the update right here. We have some metadata of the deployment, like container ID, image ID, image name, it's all here. In this case, we have just one step in the deployment. But in case you have few steps, you will see them all right here, including Walbeck, for example. Let's jump to the devices page. All right. We can see here that the -- in the Raspberry Pi device world that the version has changed accordingly to webinar app version 2, and we can also see that we have a running Python application right now on the Raspberry Pi, which is, of course, the container that is running the Python application, showing a green screen with the webinar V2 text on it. All right. From here, let's quickly jump to the remote control tool. So using it, we can remotely access the device shell and debug everything inside the device. This tool is just excellent for having a better understanding of some deep support situations. So in a few moments, we will have access to the device shell. And the reason we need to wait right now is because the device and Upswift agent running on the device is always running as the client side. So we are basically waiting the device to communicate with Upswift's servers. Let's copy secret key here and access the device. Great. So we can now debug the device, do whatever we need to make sure that our application is running as expected. In this case, let's just run docker and check what containers are running right now on the device. As you can see, we do have here a container that is currently running with the tag V2, and you can see that the uptime is 2 minutes, which is, of course, makes sense. So let's just keep the remote control open and continue to the next features at Upswift. So from here, let's jump to the remote commands tool. So in most situations, the remote commands tool will be our best friend for basic remote support just by running some commands, for example, to restart the application, maybe rebooting the device and so on. And of course, even here, we can run commands on a group of devices as we need. All right. So the last feature we are going to show very quickly is the monitoring and alerting capabilities of the platform. So using Upswift, it's possible to monitor the device resources, CPU, RAM and disc, as you can see right here, as well as the device running application. As you can see, I already monitor Python 3 application running on the device. I can define here any other application that I want to monitor. And it's all coming together to the alerting system right here where we can define different alerts. So as you can see, I have some alerts for the resources on the device and some for the processes on the device, and I can trigger them by just defining different kind of triggers. So I do have here for my Raspberry Pi an alert that said that if the device CPU is crossing 5%, send me alert about that. And as you can see, I had today an alert that crossed -- that was echoed to -- a CPU alert that was echoed to 78%. So of course, I received email about that. That's all. I will have to stop here to continue the webinar. Before that, I will say that you are all welcome to open a free account and explore Upswift features with your devices. Again, it takes just a few seconds to register and connect the device. Thank you, and back to you, Yoav.

Thank you, Amit. Great demo and great technology. Of course, there are more features to explore. But due to the timing, we cannot get any deeper at this point. So like Amit said, go ahead and feel free to open a free trial. It comes with virtualized devices as part of the trial, and you can explore further. Everything is very developer friendly. And in fact, most of the adoption of Upswift today is bottom up. We're seeing a lot of developers coming on to the platform and trialing it and adopting it for the IoT scenarios. I want to speak now about what to expect from the integration of JFrog and Upswift in the coming future. So like you just saw that this integration basically takes us to a full round of phone calls to run time with devices. So the update state, the run state, the monitor state, they all happen naturally as part of your regular DevOps pipeline. And the first thing that we are going to do in the near future is, of course, provide more out-of-the-box integration of the Upswift platform with the JFrog DevOps platform, specifically with Artifactory at first, and we're going to provide you this 360 view of your releases coming from Artifactory ending up in the run time of your IoT devices. One key element here is the integration of releases or release bundles, as we call them at JFrog. This is an easy way to consume an application. It's an [indiscernible] that defines in Artifactory both your files, your docker containers, any additional file that accompanies them as one bundle of files that you can deploy from Upswift in one go into your devices. That being said, you can and you will be able to integrate Upswift with any automation workflow. We intend to keep the open developer APIs of Upswift and the usage of it open to any kind of integration, and that's not going to change. So you can -- if you wish to start directly with Upswift, you can do it right now and you can explore it even without the JFrog platform. And of course, we're also going to invest in adding more capabilities and features to Upswift. Another type of integration that we're going to present is about the area of distribution. So we see great synergy here with Upswift. Distribution, JFrog distribution and private distribution network specifically. And the idea here is to accelerate the deployment to the devices and make it more reliable by allowing devices to pull from regional distribution nodes to which the JFrog distribution is pushing releases. So you get your update from more reliable source that is closer to the device, and it keeps a lineage to the source of the software in your DevOps platform. Yes. So just to recap, we see Upswift as a very, very synergetic piece of the DevOps tool chain that JFrog is providing. It's a natural extension to the platform. And while also serving as an open service for integration, it allows our users to achieve fully automated developer to device deployment, to gain complete visibility to this process, to monitor, to troubleshoot, you saw the remote control and to manage the secure device [indiscernible] at scale. And finally, consolidate the DevOps processes and bridge over the previously siloed workflows for device update management. And again, like Amit said, we urge you to go on and try it. And I think this is a good point to go through some Q&A from the audience. Courtney, do you want to...

Yes. Well, I just want to first -- yes, so thank you guys for a great presentation. That was wonderful. [Operator Instructions] But we do have a couple of questions that I want to kind of release and see what you guys think about it. So the first question is what verticals do you see mostly adopting edge and IoT?

Amit, do you want to take it?

Yes. Well, it's actually pretty amazing to see so many industries and verticals. So it's hard to pick one. But I will say that we see a lot of smart kiosks screens for many kind of situations and industries that just prefer to have a screen -- with a touchscreen with running application, showing what the application is doing. And it's something that I think that we'll just see. We'll see from it more and more in the couple -- in the next couple of years.

I can maybe expand on it. Actually, the amount of industries adopting and actually also requesting JFrog to extend the life cycle to IoT devices, to connected edge environment, it's very varied. So like Amit said, there are use cases of kiosks. And when we actually saw the portfolio of the Upswift customers, the innovation going and the type of things people develop and how they push the market forward. with all kind of, it can be farming, it can be even taming pets. It can be automotive industries. It can be a remote retail point of sales. So it's very diverse, right? It's -- I don't think -- like software that is everywhere, the edge software and the type of software developed for devices is cross domain. So I don't think there is any single domain that is standing out here [indiscernible].

Awesome. All right. Well, we have a second question that just came in. It's -- how can I take advantage of the IoT capabilities in the JFrog platform? And when will they be available?

So like I said, in [indiscernible] next section, you can try it out today. It's production ready. It's not a trial platform. But there is a free trial that you can try immediately. The more -- the closer integration -- so since the platform is very API-driven and very developer open, it works with the JFrog platform very easily to date. It's very decoupled connectivity, if you want to integrate both platforms. And we're going to, as I explained before, invest in more out-of-the-box features that will also allow you to gain better visibility from JFrog to Upswift back in fact to make it a very seamless experience. But yes, you can try it out today and you can adopt it today. There is no need to wait for that integration.

Wonderful. And everybody, we have our technical experts here on the line. So please don't hesitate to ask your questions. Obviously, we do have questions post on that. Amit, do you want to provide them the best way to ask those questions if they do have any after this call?

Yes. With my -- in the chat, right now. And again, feel free to reach out for any kind of technical question. We will be happy to answer.

Great. Thank you guys so much. Amit, if you want to go ahead and just put it in our, yes, team chat there, I will go ahead and provide that to the group here. Here's the -- I'll provide the contact information. It's -- because, again, I know that always post event that sometimes those questions come up and you wish you would have answered -- asked them then and you think about them later. It happens all the time. So I did go ahead and just put Amit's e-mail into the public Q&A chat for all the attendees. So if you guys have any questions after this, please don't hesitate to reach out. Again, thank you, Amit. Thank you, Yoav, for coming, taking time today to talk about this integration, in part, new partnership. I really appreciate it, and I know everyone on the call appreciates it. So I will close it out, and everybody have a great day, a great weekend, hopefully. And then other than that, we'll close it out. Thank you guys so much.

Thank you.

Thank you.

All right. Bye-bye.