Cloudflare TV

This Week in Net: The Supercloud & Developers Week 2022

Presented by João Tomé, John Graham-Cumming
Originally aired on 

Welcome to our weekly review of stories from our blog and elsewhere, from products, tools and announcements to disruptions on the Internet.

João Tomé is joined by our CTO, John Graham-Cumming. This week’s program is a special edition to talk about our Supercloud and Developer Week 2022. We go through many of the announcements of new products, tools and features.

Cloudflare’s been building the alternative to that traditional cloud into our network and our developer platform for years. The Supercloud. The term may be new, but that doesn’t mean that it’s not real. Today, we have over a million developers building on the Supercloud. Each of those developers wants to get code running on one machine and perfect it.

Check all the announcements at the Developer Week Hub !

English
Developer Week
News

Transcript (Beta)

Hello and welcome to This Week in Net, our weekly review of stories we've been writing in our Cloudflare blog, but also things affecting the Internet.

And this one is a special edition related to our Super Cloud Developer Week.

I'm João Tomé coming to you from Lisbon, Portugal.

And with me, I have as usual, our CTO, John Graham-Cumming.

Hello, John. Good morning. Actually, just good afternoon now. We're recording this just afternoon.

Good afternoon to you from Lisbon, too. And anyone who's probably seeing us in any time zone or any period.

Goodness knows. Yeah.

Yep. Yep. Some Internet time. Swatch time. Exactly. So this is the Super Cloud and Developer Week, one of Cloudflare's Innovation Weeks.

So it's all about our workers platform and things we're building.

And not only us, but a lot of interesting projects are being built with our platform.

So a lot of announcements. Where should we start?

So many things, too. Well, I mean, one of the things that strikes me is there's the millions and there's the billions.

So there's a million developers using Cloudflare.

So that was already pretty amazing. But more amazing, perhaps, is that the workers launchpad has gone to two billion dollars.

So why don't we look at that post first?

Because that's pretty crazy, right? It is. It is a crazy number.

And just a few two months ago or so, it was half of that. So it's increasing.

Well, originally we thought it was going to be, we thought we might, if we were lucky, get 250 million in funding available for startups who are building on our platform.

And then when we announced it during birthday week, we got to 1.25 billion.

And since then, we've seen a lot of other venture funds and other investors just want to get involved in it because they see the value in the platform and they see all the stuff that's being built with the platform.

So, yeah, two billion dollars.

There it is right there. Exactly. And we also have the first cohort as well of people who are joining the launchpad.

So you're starting to see some of those companies come about.

But, you know, 14 new VC partners. Look at the list of names on here.

You know, pretty, pretty amazing. And two billion dollars. So, I mean, from my perspective, obviously, we use the workers platform every day.

And so we're used to it.

But it's incredible to see companies. And here's a list of companies that are in our fall 2022 cohort who are building their thing on Cloudflare.

So they're building their technology using our platform completely.

And it's pretty incredible.

It is. And some of those are like smaller projects and others are like already big projects.

So different scales in terms of companies that are using it. Right.

Well, I mean, I think the important thing about workers is it gets used for a variety of things.

And obviously, these are companies that are, you know, taking a bet on Cloudflare as their platform.

And we hope that many of them become, you know, household names or, you know, large companies because of the speed with which they can innovate.

So I think that, you know, the platform is here. And Developer Week is slightly about that.

Right. Because we announced a whole bunch of other stuff.

But before we get into that, you know, one of the things that I think is cool is that these people are building on workers, but we also build on workers.

And there was an announcement from a team that, you know, you know very well, which is the Radar team, which is here in Lisbon and mostly.

And they announced how Cloudflare Radar 2.0, that's radar .Cloudflare.com was built.

And that was a really interesting blog post because you're getting an insight into how to build something on workers and also how Cloudflare builds things on workers.

So you can see here all of the technologies that were used for Cloudflare Radar 2.0.

And it is all on Cloudflare.

So it's a Cloudflare pages site. It uses Cloudflare. It uses functions.

It uses workers. And, you know, if you visited the website, then you would have seen, you know, just look at the list of technologies here, but, you know, on our platform.

And we have obviously, you know, lots of bits of technology in here.

And we also, if you go into this, not only does it tell you how we built it, like all the different parts we're using, but also about the API.

So that people can then, you know, remix the data themselves.

They can take the data out of our API.

And if you scroll down even further, there's actually Google collab notebooks as well.

So you can go in there and actually work with the data yourselves.

So if you are thinking about being one of those companies in the, you know, the workers launchpad, this blog post actually is really interesting because it tells you about like how we built something real on Cloudflare.

So it's well worth looking into that.

Well, we're looking into the code that we've open sourced, use collab notebooks.

If you want to go in there and actually work with some of the data that we're putting out there.

So that was a, that was a good, that was a good blog post this week.

It was, but I was experimenting, I was experimenting actually with our API and you can compare locations, which is like amazing.

It has a bunch of things that people can just see it for themselves. Yeah, go in there.

Yeah, look at that data and look at the trends in the Internet use and what's happening with traffic and stuff like that.

Incredibly, incredibly powerful to look at that.

So, so that was that, but there's a lot more in developer weeks.

So we should probably take a look at some other things. One of the, on the left there, you've got this thing about Cloudflare snippets and this is actually a really, really cool feature.

So you know, one of the, one of the things that's neat with the Cloudflare workers is you can build applications on Cloudflare and you can bet your company on it.

But you may also want a bits of code, other places in Cloudflare and snippets allows you to do that.

So we have this very powerful, you know, user interface in an API and functionality like the rules engine, which allows you to handle different sorts of traffic through Cloudflare and provide rules on top of it.

But there are sometimes you want programmability and what snippets does is it essentially allows you to take little bits of workers code and attach it in different parts of the flow as things go through Cloudflare, which gives you programmability sort of, sort of everywhere.

It also allows you to do stuff like take real legacy code, like stuff that's written in varnish in VCL and take it into Cloudflare.

And it's a nice diagram, actually, if you scroll down in here, where you see like, you know, you've got a bunch of, you know, varnish on the left there and we convert it into either configuration in Cloudflare or into snippets.

And we've done this for a while under a thing called project turpentine, but now it's actually getting really embedded into the system itself.

I'm going to do the same thing for HT access, Nginx configuration, so you can take something legacy and put it into Cloudflare pretty easily.

And, you know, if you need to, you can just write some code, right? You can write some code, like little snippets of code inside Cloudflare.

So we're bringing code to everywhere on the platform.

Exactly. And to your point, you said before one thing, that is, actually, you wrote this on Twitter this week.

One of the secrets of Cloudflare is we build on Cloudflare, right?

So the network, the global data center that we, or network that we have is one of the main reasons why we innovate so fast, right?

I think that's right. And also, if you go back to the initial blog post, which I wrote about this idea of a super cloud, and the idea of a super cloud was that, you know, if you think about what developers really would like to do, it's write their code on their machine, get it running, and then deploy it and just have it work at Internet scale, right?

And, you know, there's been all sorts of amazing things have happened over time.

EC2 came along and gave us virtual machines that we could go out and just, you know, get hold of and scale.

And there have been sort of database things in the cloud.

And then there's been something called edge computing, which was an idea of bringing computing near to the end user.

But I think there's something larger than all that, which is this idea of the super cloud, which is to say, look, give us your code and we will just run it for you.

And we will figure out where it needs to run. So it's like, oh, does it need to run actually right by the end user in that sort of edge compute style?

Or does it need to run really near a database somewhere?

And how do you make it scale over the entire network?

And so the idea of super cloud is to say, you know, it's not really about EC2 style.

It's not really about edge. You know, edge solves some problems.

EC2 solves some other problems. What you really want is just like, here's my code.

You figure out how to scale it Cloudflare. You figure out how to make it run in the right place.

How to make it run most economically, if that's what somebody wants.

In such a way that, you know, the application scales and performs incredibly well.

And obviously we have that within the workers platform, which is what we build on.

And hence how we manage to push stuff out so rapidly. So, you know, that was sort of the motivator for the week was super cloud.

And then within the week, there was a bunch of other stuff announced.

True, just one highlight here.

It's also about flexibility, but also costs. So having a total ecosystem reduced costs for sure, right?

Well, I mean, the other thing is that, you know, one of the problems with the EC2 style of things is that you have to figure out how you're going to scale your system.

And often what you do is you have sort of hot standby machines that you're paying for.

And it's also kind of inefficient because you have to have entire machines waiting.

Whereas you could actually just have your code waiting.

And so I think it becomes just, you know, it's more environmentally friendly.

It's more flexible. It's less work because you're not thinking about the scaling of the application.

And that's what super cloud is about.

I like the word. I like the word because it's, and I like that we, it's a word that wasn't there at the beginning.

And then it was, so it was like a few things.

And then, hey, this is like a super cloud. It is. Yeah. It's a super cloud.

Exactly. Well, I think it was, I think one of the problems is those, you know, there's been this like edge compute thing.

It's like the next evolution of cloud.

And it's like, yeah, I mean, I think edge compute is important. And I think you probably want to have some edge in your thing.

But I think actually what you want is the platform to figure out where the code runs.

And to that end, we should probably talk about D1, the database.

So D1, which we announced some time ago is now, here it is.

And it is now in open alpha. So that means you can then go in there and spin up a database on Cloudflare just like that.

And in fact, it's got, you know, there's been tremendous demand for this.

And now it is available to anybody.

And what we're seeing is, you know, very large numbers of databases getting created very, very quickly because, you know, traditional databases are very, very useful for building applications.

And so what you really want is not to have to worry about how to manage one, but have somebody handle it for you.

And that's what Cloudflare is doing here with D1. And I think one of the important things is to think that, you know, D1 is just the first in a number of database things that we will be relaunching.

And we previously launched key value store, of course, with workers KB.

We previously launched our object store with R2.

And, you know, R2 and D1 are not the end of this. There'll be more things, but you can use it today.

It is in alpha. We still think there's some kinks to be worked out.

We've talked about in the blog post what the restrictions are, but it is there and it is part of the super cloud experience, which is to have a database that just scales up.

There's another blog post actually.

And we want people feedback, right? In terms of making it better. Absolutely.

And there's also a blog post about actually building something. So this making static sites is the next blog post there.

Making static sites dynamic with D1, where what you'll see here is there's actually a really nice worked example of saying, right, I'm going to add functionality to a static website using D1.

And this goes through how it's done. So you see like the database getting created.

And, you know, in this case, it's adding a comments to a static blog. And so it actually, you know, goes through here and shows how to do that using, you know, using the database and what the commands are.

And, you know, it's super simple. And we worry about how to scale that.

We worry about where to execute it. We worry about where the code should be in the database.

So hopefully we'll see lots more things like this get built over time.

Exactly. It's all about making it simple, more easy.

You won't need like so much trouble to build stuff. So it's always helpful there.

Yep. And we're adding stuff. Where do you want to go next? Well, I mean, good question, actually, because there's so much this week.

You know, speaking about static websites, let's talk about pages functions.

So, you know, pages, wonderful static website service from us, put your stuff in GitLab, GitHub or GitHub or GitLab or wherever you want and then push it and we will build it and we will publish it.

And I use that extensively for my own stuff. Just like we were doing with snippets, it's always often useful to have little bits of code around the place.

So snippets is all about how Cloudflare's own functionality gets modified by code and pages functions are all about how you can add little bits of code to a static website.

And so in here you'll see it just goes in a functions folder and we will execute them against the static website.

So again, Wrangler, you can go in there, you can see how the thing is operating and you can build, you know, really rich applications with this.

And so, you know, we're seeing this already happening with people building things on Cloudflare pages with functions.

And there are some other examples this week actually from the team around the architecture of front-end development and, you know, applications that are used by end users with server -side rendering and also micro front-ends.

Those, both of those are really quite deep dives into how you would do that.

So the micro front-ends sort of movement we've talked about a little bit before on the blog, but this is about how to do it, how to do incremental adoption using Cloudflare Workers.

And again, lots of detail on here about how to do it with the platform if you're a developer.

And then the other blog post, which is about server -side rendering is about how you can use our platform to do server-side rendering.

It's just above, I think there it is. And, you know, again, now you can do that using pages functions.

So this is all coming together into, or it has come together into a service where you can build a static website.

You can add functionality to it. You can use micro front-ends for bits of it.

You can use functions to do server-side rendering. So I think, you know, if you're building something, there's a lot of technology here and it is super quick on our platform.

Exactly. And again, it's always, all of this is feeds on the global super cloud.

Yeah, this is all part of the super cloud. Exactly. The low latency, 275 data centers across the globe.

So it's all about getting more things into what people can do using our network, really.

Exactly. Exactly. I mean, there's just sort of a tremendous amount of stuff people are building on the network.

And then if you go back to that at the beginning of the developer thing, I mean, for larger scale applications, we talked about queues.

So queues functionality is there and you can build an application of any size on our platform.

I think this is a key piece of functionality for unlocking, building very large applications on Cloudflare.

So here it is. It's open beta. So go use it today. Exactly.

And we have in all of these blog posts, a lot of examples in terms of even how to use code, how to make the better, the best of some of these new functions and functionalities.

Absolutely. Some are beta, some are already open available, and we have also an alpha one.

So different stages. I mean, it wouldn't, it certainly wouldn't be a developer week if there wasn't lots of cool functionality in there and lots of stuff to actually try out.

Because, you know, the one thing developers want to do is to, you know, at least get Hello World working and then start to imagine what they can build with it.

Exactly. And it's a really big community.

So it's getting bigger. So it's kind of amazing to see the growth. And we have the developer Discord channel that more and more people are using too.

And then we did a bunch of work on our API.

So our API has been switched to use OpenAPI. And so, you know, that's, that's making a big difference to us.

And it's made our documentation better and clearer.

So you can see there's a bunch of blog posts about how we did the OpenAPI stuff.

And so I'm really, I'm really glad we did that work because the API is an important part of what we do, right?

Yeah. The one you just clicked on is about the winter CGs.

That's about the working group that's trying to standardize how the sorts of development environments that we have and that are similar, you know, and others that are related.

So Node, Deno, BUN, et cetera.

They have a standard platform API because what we don't want to do is to lock people in, in a silly way where it's like, well, you can't go anywhere else because you're using all these magic beans that are on Cloudflare.

You want to have maximum, you know, compatibility with other, with other platforms.

So there's a big effort with the winter CG.

And this is just about our compliance with those standards and what's working and what's new and what's not.

So there's a lot of big improvements here.

A lot of rewriting that I showed out aside to make sure that we are compatible with, with these common standards.

We actually got to use other, yeah, you can use other platforms, you know, and when we started with workers, I mean, you know, why is it called workers?

Well, in part, because we started out with service workers, which is the browser standard.

And we wanted it to be something that is, you know, very common for people to be able to, you know, they can use it.

There's documentation about it.

It's not something that, you know, we entirely invented ourselves.

So that's good. And then, oh gosh, there's so many other things.

The one that's on screen there about automated browser isolation, that's actually really, really cool.

Just that you, if you need to, you can run, there's actually some code in that one where you can run an entire browser from Cloudflare workers.

So for example, if you want to take a screenshot or you want to do an AB kind of test, we want to see how things look in different ways.

You can actually automate an entire browser, you know, because we have the browser isolation product, which allows us to run a browser for you.

What we've done is we've basically integrated Puppeteer into the workers platform.

So here, look, this is all the code you need right here.

You just scroll past it to start a browser, take a screenshot and store it in R2.

I mean, it's pretty incredible, right? So you can just do that.

And it's short. It's really short. It's really simple. Right. So, and so this is using our browser isolation technology.

And that is now part of the entire platform.

So another great announcement this week. And so, so many more things this week in developer week, you know, I think, you know, and as we say, if we go right back to the beginning, a million developers, $2 billion in funding for people building on our platform, because people are building all manner of amazing things on the platform.

For sure. Announcing also R2 super slurper.

Super slurper. Okay. So the idea is like, look, if you think about R2, what's R2?

It's S3 without the egress fees, right? That's why it's one letter less than one number less.

It's got less than, it's got less than S3 and less cost, right?

And one of the things you want to do is you want to pull your stuff from S3 or some other S3 compatible thing into super slurper it's into R2 because it is cheaper.

And we had a product for doing these kinds of migrations. We have products called the R2 super slurper.

And the idea is you can, you know, we can, you can set it up.

So it slurps all the data out of S3. If you want to do a, you know, a lift and shift to the entire thing, or you can do it sip by sip.

So actually as, um, as a, as an object is needed, go get it from S3.

If it's not already in R2 and then insert it.

So the super slurper is designed to help you with those kind of kind of migrations and private beta now, and we'll be building that out over time, over time so that people can move their objects into a storage, which doesn't, it doesn't lock them in basically.

Right. I mean, the thing about egress fees is it's a tax on getting stuff out.

This is a way of helping you get them out one more time without paying any further times.

Right. So pay this once, and then you are free of egress fees.

And I think this is really important for opening up, you know, the use of, of object stores in a cost effective way.

And also frankly, giving people freedom to move around because guess what?

If you put everything in R2 and we have no egress fees and you don't like Cloudflare R2 for some reason, you've decided to move to some other cloud, well, you know what?

You can do it. You're not being artificially locked in by egress fees.

And I think that's an important part of how Cloudflare thinks about applications.

And you shouldn't be locked in artificially.

You should like our platform. Exactly. And that control sometimes is really damaging for a company because a company sometimes starts small and then it's got bigger and then they, oh my God, we're closed into this thing that we joined.

So having like more freedom there really helps, right? Yes, absolutely.

You, you, you know, and you don't want to be locked into, into your platform in a silly way.

So we'll still have a few minutes.

Do you want to go through another one of all of these amazing, there's so many, right?

There's so many things. So, I mean, there's a couple of things I hear that I think are kind of, you know, fun, which is if you just scroll back up again, there's a thing about, yeah, the thing you just hired today is Zata workers.

So there's a bunch of customer stories here about people are building on our platform, well worth reading what people are doing, you know, how they use our platform in case of Zata in terms of how to keep client side or not have client side secrets.

If you want to do database access directly from the client side.

In fact, you do it through, through the workers platform. And there's a bunch of examples this week for that.

But the other one, if you just go back to the developer platform thing, the other one I was going to, I was going to mention is that if you, if you scroll it, yeah, scroll it back down again, there's right there.

It's about authentication with mutual TLS. And that's actually kind of a fun story, which is that, you know, in order to be really, really sure that it's your application, which is contacting, being contacted by Cloudflare and the other way around, then we can use mutual TLS.

So that, that functionality is coming.

And so that as, you know, as requests come from us, you can be assured that they are coming from our platform.

And it's, you know, it seems like a relatively small thing, but actually it gives people a lot of assurance that it's their worker that's calling, you know, is calling their, their object on whatever backend they're using.

And obviously we can secure the other direction, right? Which is to authenticate that we are talking to their genuine, their genuine backend.

And so, you know, here's a piece of functionality, which is useful for customers, but also is useful for us, right?

Because we want to be able to do things internally too, that are authenticated.

So this gives us service to service authentication.

So this is also something really important in terms of cybersecurity, right?

In terms of being protected. Yeah. Well, I mean, you want to be sure that the, you know, if you're talking to some backend somewhere that you are talking to the right one and that the backend wants to be sure that it's Cloudflare that's doing it, not someone else.

So yeah. So mutual TLS for workers is coming soon.

And if we have a moment, if we scroll all the way down to the bottom, there is a blog post about using Terraform.

So Terraform is a nice open source project, which many people use for managing a configuration of large things.

And in our case, we use Terraform to manage Cloudflare.

So we have a Terraform provider, which allows our customers to manage Cloudflare, their Cloudflare account through Terraform.

And we use it as well. And we use it to manage how we use Cloudflare because Cloudflare of course is built on Cloudflare and we have to manage our configuration.

And this blog post talks about how we thought about it and how we use the thing called Atlantis, which is a sort of related tool to help automate how we use Terraform.

So interesting if you are in the business of managing a large piece of infrastructure, this gives you some insight into what we do internally to manage our large infrastructure and our use of Cloudflare itself.

So do take a look at that because it's based on the real experience of using Terraform at scale.

Exactly. And there's real advantages of us dogfooding what we launch because it's the first criteria of getting bugs, understanding how to make things work better, right?

And then there's the feedback of customers and beta and all that.

But us dogfooding is like a first tier of getting things really well done.

Yeah. I mean, obviously we want to dogfood our product. We use our products to protect Cloudflare, to accelerate Cloudflare.

And we want to use them. And I think it's just important.

If we weren't using them, then well, that would be a pity, right?

We should be using our products. We shouldn't be relying on something else if we provide something.

So for our weekly company meeting, we use our own stream product to stream the video.

Exactly. So I think it's important that we know how our stuff operates.

And sometimes we have more than 1,000 people there. So it's good also to see things that we're doing.

I'm showing Altrind here. Yeah. I mean, we've done some pretty big events where we've been streaming concerts and stuff like that.

And in order to be sure we can stream a famous pop singer's concert, we can start at least by knowing that we can stream our internal meetings and then scale up and scale up and scale up.

And so we definitely wouldn't have used something else. We're almost out of time.

Anything to wrap up things? We're not over, right? I mean, we're recording this.

And in fact, we're about an hour and a half away from a whole bunch of announcements that are today.

That's true. One of those announcements. So the ones for today, what would be the main ones that you want to have?

Well, I think the network performance update.

Every time we do an innovation, we try to do an update on our network performance because we are out there measuring our network performance, trying to make sure it operates correctly, make sure it's the fastest, make sure our compute platform is the fastest.

And we've got an update on the measurements we've been doing around that.

So that's coming out today. But there's a bunch of other stuff about TypeScript, about improving support for TypeScript is coming out today.

So more and more developer stuff today. So yeah, don't miss Friday as well.

People can find all of this in blog.Cloudflare.com. And of course, we also have a dedicated page that we were showing actually before in our Cloudflare page.

And next week, it's Thanksgiving and like Friday. And we're also monitoring a little bit of e-commerce websites trends in terms of Internet trends in general.

Yeah, well, you and I won't be doing Thanksgiving since we're in Europe.

So we'll be back next Friday with this week in net. And there'll be a bunch more stuff next week because we've got, I think, a lot of blog posts about quite nerdy networking topics.

So we would expect some deep dives into networking stuff over the next week.

Next week. Perfect. Thank you, John. That's a wrap. Cheers.

Thumbnail image for video "This Week in Net"

This Week in Net
Tune in for weekly updates on the latest news at Cloudflare and across the Internet. Check back regularly for updates. Also available as an audio podcast!
Watch more episodes