Follow me:
Listen on:

Day Two Cloud 180: Understanding AWS EC2 At The Edge

Episode 180

Play episode

On today’s Day Two Cloud podcast, we speak with Jan Hofmeyr, a vice president within Amazon Web Services (AWS). His group focuses on Amazon EC2 Edge, including AWS Outposts. This show was recorded at AWS re:Invent 2022 in Las Vegas.

We discuss:

  • EC2 at the edge
  • What goes on in an AWS Outpost box
  • Standing up an AWS Outpost box in a local rack
  • What an AWS Local Zone really is
  • Connecting an AWS Local Zone to the AWS cloud
  • More

Show Links:

AWS Local Zones – AWS

AWS for the Edge -AWS

AWS Outposts Family – AWS

AWS Private 5G – AWS



[00:00:10.970] – Ethan
Welcome to day two. Cloud. In today’s episode, we are interviewing Jan Hofmeyer. Jan is a vice president within Amazon Web Services. His group focuses on EC Two. Specifically, EC Two functioning at the Edge. So we’re talking talking local zones, AWS, outposts, these sorts of things. Ned and I recorded the show at AWS Reinvent 2022 in Vegas. Some media folks were kind enough to set us up with Jan to have this conversation in the discussion. Ned and I asked Jan all kinds of questions about EC. Two at the edge. What’s really happening in an Outpost box, how you order one, what it means when you’ve got an Outpost box standing up in one of your racks, what a local zone really is. How does AWS decide to stand up a new local zone? What is the plumbing for AWS like going into a local zone? Is it as good as what happens in a region? Jan answered all of our questions, and it was totally unscripted. We walked in having really no idea what all Jan was ready to talk about. We just went in and started asking questions, and he started giving us answers. It was a lot of fun.

[00:01:14.180] – Ethan
So enjoy this conversation with Jan Hofmeier, vice president at AWS for EC Two at the Edge.

[00:01:21.370] – Ned
Jan hoffmeyer. Welcome to day two, Cloud. We’re very excited to have you here today on the show. Can you please tell the good folks out there a little bit about yourself and what you do for AWS?

[00:01:32.190] – Jan
Great to be here with both of you. So I’m with the AWS EC two team. And EC two is the cloud. Compute for Amazon. And specifically my focus is around the easy to edge area. So before you even ask me what is the age the Edge is at AWS, we have our big regions, and our regions are our large data centers. Although locations, we have 30 of them across the world. Each of these regions have 97 or not each, but these regions represents 97 Availability Zones. So that Availability Zone is for a given region. There’s typically three or more or four different Availability Zones. Each of those are independent data centers. That’s where we run the big workloads. So think of that. That’s AWS, right?

[00:02:22.500] – Ethan
That’s easy too, that everybody knows and.

[00:02:24.060] – Jan
Uses all of our services, over 200 services. That’s Amazon. Right. AWS, the Edge is when we take the Cloud, the AWS Cloud, outside of the region and closer to our users outposts. Well, the first stop is actually Local Zones.

[00:02:40.090] – Ethan

[00:02:40.950] – Jan
So Local Zones is where we still have AWS run facilities that runs a smaller instance of the AWS Cloud, closer to metropolitan areas and the Edge. And a good example of where that’s useful is the Dish. For instance, when Dish deployed their 5G network in the US. They can run some of the non latency sensitive applications in the region. And then things are latency sensitive, and the 5G network that comes close to a tower that they can run in a local zone that’s in the metro, close to the tower. So that’s the local zones. And then we go closer and we go on Prem. And for on Prem, whether it’s your data center, your office, your factory, your retail store, we have outpost, and we make outpost available in two forms. One is a 42 U rack. So for the listeners out there, that’s like a big sub Zero refrigerator. Yeah, it’s a large Barack with equipment. But we also have a one unit, two server, just an individual server. And so that’s the edge.

[00:03:45.260] – Ethan

[00:03:45.910] – Ned
Okay. How does Snowball fit into the Edge picture? Does that fall under your umbrella as well? Because I know there’s, like, Snowball and there’s Snowball Edge that has some compute.

[00:03:54.670] – Jan
Bundled in well, my colleague Wayne Dusso, he runs the Snow family of devices.

[00:04:02.440] – Ned

[00:04:02.940] – Jan
And the Snow family differentiates itself from outpost that it can run disconnected. So a Snow device, you can imagine, can run in an area where you don’t have always connectivity to the cloud.

[00:04:14.360] – Ned

[00:04:15.270] – Jan
And it’s typically used. The most extreme case was we send one up to the International Space Station and you would think, is that a publicity stunt? Why would you do that? And there’s so much data generated there that their ability to just load it up into a device. With the next vehicle coming down, they can get the device connected. All the data goes into AWS, and their scientists immediately have access to it. So it’s actually a really interesting use case that NASA is using it for bringing data down, large amounts of data down from the space station. So, yes, it’s in the disconnected world. So that’s where snow looks.

[00:04:50.780] – Ned
Talk about being bandwidth constrained.

[00:04:52.570] – Jan

[00:04:53.090] – Ned
The ISS, that makes a lot of sense from a storage perspective. So tell me a little bit more about local zones. How do those differentiate from, like, an availability zone per se?

[00:05:05.330] – Jan
So the idea of a local zone is how do you get closer to where the users are? And it’s really driven by two things. One is latency. A great example of latency that I’m sure many of your listeners will appreciate is gaming, where latency is really important. So what a lot of your gaming companies like Epic will do is they will deploy in local zones. That gives them a much better latency experience. So that will be an example of latency. The other one is data residency. This is where there’s regulations that says the data must be in a country or in a state. They cannot even leave the state. So a good example of in the year in the US. Of regulations where the data has to stay in the state is real time gaming. So companies like FanDuel, okay, the data has to stay within the state. Then they can run in these local zones or even in outpost in the state and then be in compliance with the regulations.

[00:06:03.770] – Ethan
Does that help with GDPR as well?

[00:06:05.930] – Jan
It does, but GDPR we support across the entire Amazon continuum, right from the regions all the way down. Because at the end of the day, it’s really the customer that controls where the data lives. It’s the customer who controls who have access to the data. So GDPR compliance is across the entire.

[00:06:23.790] – Ned
Cloud because it’s a smaller facility, it’s still managed by AWS. What do I have from a resiliency perspective for that Local Zone? Because I know best practices for an application running in a region would be to put it across a couple Availability Zones in case a zone were to go down. What do I do at the local zone level?

[00:06:42.660] – Jan
So with the Local Zone, what we see our customers do is they run their application in the Local Zone and they use the region as the failover if something do happen with the Local Zone. Local Zones are fully managed data centers with the same resiliency as we bought our Availability Zones in the region. Right. So it’s the same level of resiliency that we built there, but if something had to happen with the Local Zone, they will fall back to a region.

[00:07:07.230] – Ned
Okay, that makes sense.

[00:07:08.460] – Ethan
Now, the big play for a Local Zone, you were talking about latency moving those workloads closer to where they need to be processed. Can I get all the network services that I would get Direct Connect circuits and this kind of thing into my Local Zone?

[00:07:20.980] – Jan
Absolutely. So we support many of the same networking service. Mesh AWS you mentioned AWS direct Connect. That’s a critical service that we support in the local zone and just for the readers, that gives our customers the ability to have a private, dedicated connection into us from whatever facility they come from, whether it’s their data center or their offices, into us. Absolutely.

[00:07:43.800] – Ethan
Okay, now we’re a database reinvent 2022. Your world is easy to an edge, as we’ve decided here. So what announcements have come out this week that you want to share?

[00:07:54.790] – Jan
The first one. I’ll start with the local zone. We are expanding Local Zones into international markets. We’ve announced we will go to 30 international markets. We’ve done already eight of them this year. We’ll continue to increase that number. Why? That is really exciting. Is there’s many countries back to the question about data residency. There’s many countries where we don’t have a region where we are putting Local zones, and that really gives customers in those countries the option of running both low latency in those countries as well as being compliant with data residency requirements. That for us is exciting one with Local Zones.

[00:08:32.520] – Ethan
So architecturally AWS, AWS, you’ll keep your Regional Zones pretty much where they are now. It would be a pretty big undertaking to make a new region, but Local Zones are smaller and easier to deploy.

[00:08:42.400] – Jan
Maybe. Yeah, you know, the beautiful thing about local zones is we can start super small and I mean down to five or six racks.

[00:08:51.140] – Ned

[00:08:51.710] – Jan
And we can grow it up to hundreds of racks. We can make it really big. Thousands of racks. So it was designed in a way that can start very small. I’ll share a little more of insight. What runs inside of a local zone is actually outpost.

[00:09:06.510] – Ethan
Okay, that makes sense.

[00:09:09.410] – Jan
We can start really small and we can scale up, and you don’t have.

[00:09:12.700] – Ethan
To build a facility for that. Then you could partner with some other data center in the area and use some of their space.

[00:09:17.640] – Jan
Yeah, local zone is AWS facilities.

[00:09:21.560] – Ethan
It is.

[00:09:22.040] – Jan
Okay, so it’s us. It’s fully managed.

[00:09:25.080] – Ethan
So you’ll build something even if it’s just five or six racks?

[00:09:27.970] – Jan
Absolutely. No, we don’t bolt it for just five or six rack. We can start that small. The facility we would build typically can grow much larger.

[00:09:35.490] – Ethan
The expectation of expansion, sure.

[00:09:37.260] – Jan
But we can go into a city or into an area, start small and grow up. We don’t start that small. We know we start a little bit larger than five, but we could start that small.

[00:09:47.020] – Ned
Right. It’s a little different than some of the other edge services that I know they’re out there that kind of do this pod based architecture where they’ll drop a pod in metropolitan area and that has a maximum of, say, maybe ten racks in it. It’s easier for them to deploy, but you lose some of the expandability there.

[00:10:04.440] – Jan
No, absolutely.

[00:10:05.240] – Ethan
But it’s still tricky though, if you’re going to stand up a new facility, because you got to get network connectivity, multiple network connectivity into that facility.

[00:10:12.630] – Jan
So when we decide on a new location, there’s three, two big things that has to happen. First is we need to find space. Right. So you need to find space. We have power and ability to do it, and then your hearts and rec. The next thing is we have to bring our core network to that facility and not just one of the links. We have to have redundancy coming into bringing network to these facilities. That is a big amount of taking every time we launch one of these new facilities.

[00:10:39.280] – Ethan
Okay, so you’re picking metro regions that are going to give you the ability to stand up those facilities. But then also, is it customer demand? How do you decide where to put up a new local zone?

[00:10:49.060] – Jan
It’s 100% customer demand.

[00:10:50.730] – Ethan

[00:10:51.310] – Jan
Right. And you’ve heard so many times here at Reinvent that, you know, we are just being so obsessed that’s what we listen to. We listen to our customers and we hear to what’s their problems and that’s what we responded. So a good example would be with Dish. Dish came out. We didn’t have 17 local zones in the US. We had one, it was in La. And we listened to the requirement and the need for them to have more facilities closer to the towers. And we responded to that. So we actually respond to customers and their needs.

[00:11:26.830] – Ned
Right. So in that situation, Dish had the towers, but they don’t have the room in the towers to put all the gear that they might want for compute and processing. So they’re leveraging your local zones to do that additional processing. I know 5G has the capability to do local breakout. Correct?

[00:11:42.380] – Jan
That is correct. So with five G, the reason why I’m so excited about 5G, there’s a lot of fanfare about all the cool stuff it will do as a technologist and we are really excited. It’s the very first network ever designed and architected to be Cloud native. So what does that mean? Instead of having a network defined and said, here’s the entry point and here’s the exit point, they took every function in 5G, they broke it out into individual components. They created APIs and you can now deploy them individually. So take the dish example. Now, for the first time, Dish can say this one function is not latency sensitive, it’s a control function and I can deploy it in the region, right? And here’s a function that sits in the data plane and latency is highly latent sensitive and I can deploy that in a local zone. So it really allows you to break that whole 5G monolith almost, that you have to deploy in one place and break it up and run the function with workload in the right location.

[00:12:45.020] – Ethan
It’s interesting, the Telco story, they’ve been working on this for years now, being able to port those functions over to COTS. And the story was always about cost savings and being able to spin up and spin down some flexibility there with those services, being able to move them around for latency purposes. That’s a great story. I hadn’t heard that one before, but.

[00:13:04.280] – Jan
It’S fascinating and it shows. So one of the things, if I take a little bit of a step back, the team’s main objective is how do I bring the AWS Cloud to these locations closer to the user, right? Whether it’s a local zone or an outpost. And every time I put it down, I’m literally putting a piece of AWS Cloud down. It’s the exact same API, all the exact same tools are available. And from a developer perspective, it just looks like AWS Cloud. It looks no different. I’ll give you an example of when a developer goes into the AWS console and they say, I want to deploy in US east One, they pick the availability zone they want to deploy and then they deploy their code. Literally. The next option, if you have Outpost, it shows up as just another location and the exact same code you just deployed there, you can deploy it to the outpost. So it’s truly AWS Cloud wanting in these locations all the way to Outpost, which runs onprem but that’s so many.

[00:14:05.230] – Ned
Questions that I have. So I know one of the difficulties with hardware that’s installed on premises is typically you’re responsible for updating and patching that hardware and software, which means that your API versions might be lagging behind what is running in the cloud, because it’s easier to update the cloud. Well, you’re not updating it. AWS is updating it. How does API updates and versioning work with not only local zones, but more specifically the outposts that are deployed on a customer’s premises?

[00:14:36.650] – Jan
The very first mind shift that needs to happen with Outpost is you’re not getting a server, you’re not getting a rack, you’re getting AWS Cloud.

[00:14:47.130] – Ned

[00:14:48.040] – Jan
Right. So it’s truly for folks who run today data centers, and they get a server, they install the server, they have to provision or they have to install the software. They have to make sure that all the code wants. With Alpos. You’re getting AWS, Cloud. What you do with Alpos, you connect it to power, you connect it to the Internet, it shows up. And for us as organization, that Outpost server looks no different than the server in the region.

[00:15:13.030] – Ned

[00:15:13.600] – Jan
When I update the region, I update the Outpost server on the on premise. I update the server in the local zone. So all the same mass scale operations that we have running all these regions, the exact same tools and system automation runs local zones and Outpost.

[00:15:31.910] – Ethan
So if I deploy Outposts as an AWS customer, is that my stack of compute or is that part of the AWS cloud? And other folks, other tenants could exist on the My Outpost.

[00:15:42.980] – Jan
So outpost is your compute? Okay, so what happens? Typically, a customer go into the console and they would specify what they’re looking for. So I want an M five compute instance, right? And I want EBS block storage and so much ABS block storage. And they think about what they want to run. They want to run container services, whether it’s EKS or ECS, what networking services. They could even run Ace three on Outpost, right? And how much of a three capacity they want. So they pick what they want and then recreate that rack for them. We’ll ship the rack out to the customer, they power it up, and then when it provisions, reprovision the rack. And then when they go to the console, there’s everything they ask for as if it was in the region, but.

[00:16:28.960] – Ned
It’S dedicated to them during the update process. I know typically you would probably evacuate a host in order to update the software on that host. You’re running with limited capacity on this outpost. So how are you do you evacuate the hosts or can I control when the update happens in case there’s going to be a maintenance period?

[00:16:49.550] – Jan
So we have a super cool technology. It’s called nitro. And so what Nitro is, is many years ago, we realized that to run at scale, you can’t vacate workloads, but it would be a horrible operational overhead. I think every time I have to make an update, I have to move people around. So we created, we took all of the management of the server, whether it’s the compute, the storage or the network, and we actually moved it off the store of the compute. We’re running it in a dedicated separate hardware. That hardware runs all of the systems. So if you look at an outpost and any server in our regions, the compute storage network on the server is 100% available to the workload. We don’t want there, we want completely separate. That allows us to do almost all updates completely live and seamless to the workload.

[00:17:46.390] – Ned
I see. So if you need to patch the hypervisor, you can update it without touching.

[00:17:49.840] – Jan
The word, without touching the workplace running.

[00:17:52.600] – Ned
Nice. It’s magic.

[00:17:57.190] – Jan
In my previous careers, you work with dictators Andrews and first of all you’re voting nightmare to get everything patched and everything running and you have all kinds of stuff running and that’s just the nightmare on zone. And then the worst thing is day zero security gets announced and you have to patch right on our side. In most cases we’re aware of hosting even before it becomes available. We patch all of our regions, all of our local zones, all of our outpost. So here our customers really benefits from the fact that that outpost is the same as the region and as we patch, we patch everywhere.

[00:18:38.050] – Ned
Well, that is slick. Looking at other edge platforms that I’ve worked with in the past, that was not the case.

[00:18:45.500] – Jan
No, I’ve had that same experience. I have to tell you, it’s something with magic.

[00:18:55.290] – Ethan
Now you mentioned earlier that more local zones are being deployed in international locations. Was part of the announcement like where you were going to be doing those deployments?

[00:19:04.340] – Jan
Yeah, we had four actually announced just before reinvent in Helsinki muscat, I’m going to miss a few and more coming before the end of the year. And you’ll continue to see us roll out these local zones throughout the year. It’s a critical part as we see these local zones land in all of these different areas. It’s just amazing. This is part of the magic for me about the cloud is we have some workloads in mind and some intent and then you make it available and then people do things you’ve never even thought about. So we’re very excited to see what people are going to do with local zones.

[00:19:46.090] – Ned
To go back to an earlier question, Ethan, I think you were starting to ask and then we got sidetracked. Always happens in terms of services that are available on outposts because I think about the 200 plus services that AWS has, not all of them are available in all regions when they’re launched or later. What is available generally on outposts and does it depend on how big of a deployment I have?

[00:20:14.930] – Jan
First I’ll start with your last part of your question. It doesn’t depend on the size.

[00:20:18.620] – Ned

[00:20:19.020] – Jan
So all outputs, whether it’s one rack or multiple racks, it’s the same services. We start off with the EC, two instances, which again, like as I said, the same instances once in the region. So the instance of the very first. Next comes EBS. And Block Storage is just such a fundamental service to have block storage. Then we add some of the databases. So we have RDS, ElastiCache s three on Outpost. Then we have the container services ECS and EKS. And we just announced a really cool version of EKS. It’s called local cluster support for EKS on outpost. And what Local Cluster does is before if you lose connectivity to the region from the outpost, your Kubernetes containers will stop because you lose connectivity to the cluster. So we introduced local cluster. We can run your entire EKS cluster locally on the outpost. So even if you lose connectivity, you can continue to manage your clusters locally on the outpost. So there’s a local interface where you can then manage that and continue operating. So when the connection comes back, there’s fiber cuts, there’s weather events, many things happen, but it didn’t survive that. So that was a really important one for us.

[00:21:40.190] – Jan
And we are continuously focusing on how do you make Outpost static destable static stable, meaning that it can disconnect and continue to operate. And then for any other services, we mentioned the network service Mesh Direct, connect directly on the outputs. And then for all the other services, users use it in the region or connect to the region services to take advantage. Cloud Watch, Cloud Trail, and any of the other services they can leverage.

[00:22:10.130] – Ethan
But the outpost is built, I think you said, based on customer spec, too. So you load up the services they are asking for.

[00:22:17.480] – Jan
Yeah, some customers might say, I don’t need this one. I don’t need s three, for instance. And then we’ll configure without s. Three. Or they want s three. They don’t want. They can literally configure what they want and we send them a very specific configuration.

[00:22:32.710] – Ethan
Okay, but is the full menu of Amazon services available on apples if I want?

[00:22:37.750] – Jan
It’s a subset, and it’s really focused on the edge use cases that are very much compute, because the reason why you run on the edge is really is because of latency and it’s really compute focused. So that’s the majority of the use cases today. But we also support local data solution. So if you have a big Nas solution running on Prem, you can connect your outputs to the existing data stores you have.

[00:23:04.090] – Ned

[00:23:04.800] – Jan
And it just becomes the compute area for you.

[00:23:08.010] – Ned
I was going to ask about that because typically with hyper converged solutions, the storage scales with the number of compute nodes you have. But if you need additional storage, how would you do that with Outposts and.

[00:23:18.290] – Jan
The answer, you can connect to it. The other one I want to just we introduced the Outpost Server. So let’s just imagine you can imagine AWS Cloud running in a rack. Now you have AWS Cloud running in a single server. And the nice thing about the server, and I’m super excited about the server, is this now allows us to go to many, many locations. Typically rack customers. They have one, two data centers, maybe up to hundreds of sites. But that’s kind of the order, man. With the server. We were thinking, this is going to go to like, retail stores. We have thousands and thousands of locations. So the first thing we change with the server, we introduce with the server is it’s fully self installed, so you don’t need an Amazon technician to come on site and install it for you.

[00:24:04.620] – Ethan
You just answered another question. You’re shipping me a box of Amazon.

[00:24:07.530] – Jan
Let’s say I’m shipping you the server. You come in, you order through the API, say, on the server, here’s the address. We deliver the server at the address. You open it up, you put it in, you connect it to the power, you connect it to the Internet, and then we immediately will detect it and we’ll provision it to whatever you ask for. If you said, I want an M five with this bug memory, it comes up and in your console, that server shows up as an instance.

[00:24:32.840] – Ethan
You ship it to a customer so that when it’s you would plug it into the Internet. It’s going to come online, it’s going to phone home, I’m assuming. And then, you’ll know, we ship that to this customer. This is part of their AWS, and.

[00:24:42.520] – Jan
I provision it for them on their account. Unless you’re in their account, it shows up and there it is, and they can use it.

[00:24:47.800] – Ethan
And so the one server is something I buy outright or I lease it.

[00:24:52.050] – Jan
Or no, in, in all cases, you don’t buy the hardware.

[00:24:55.570] – Ned

[00:24:56.130] – Jan
You to a point. You lease the capacity. Okay. So think of it as a reserve instance. You can do a one year reserve instance, or you can do a three year reserve instance. So that’s the model.

[00:25:09.300] – Ned
Okay, so if I wanted an Outpost server for my home lab because why not?

[00:25:13.620] – Jan

[00:25:14.290] – Ned
I get a year upfront and get a particular rate.

[00:25:17.900] – Jan
You can either pay upfront or you can pay as you go, right, for a year. And you have a choice there.

[00:25:23.020] – Ned
Well, obviously I pay upfront.

[00:25:24.430] – Jan
Yeah, of course you wouldn’t.

[00:25:28.070] – Ned
I like to be heard fiscally frugal.

[00:25:33.930] – Ethan
So those servers come in multiple sizes, or is it just one storage Ram CPU kind of a package?

[00:25:39.360] – Jan
No, we have no. So we have the two U server, which is an intel based server. It comes in various sizes in terms of the number of compute storage you need. So you can configure it. Same where you do instance configuration in the region. We also have a one U, which runs the Graviton. If you want to run a Graviton workload.

[00:26:02.750] – Ned
Right. That’s the arm based.

[00:26:03.800] – Jan
That’s the arm based workload. And the nice thing about the Arm lane is you have so much better price, performance and power consumption. So the Arm runs much lower power. So if you start thinking putting servers in retail stores, you don’t have the type of power you see in Data Center. If you start stacking them up, I think that the cost, performance and the power is a big advantage. Yeah.

[00:26:28.110] – Ned
If it’s pulling something that’s similar to what a decent desktop would pull, that’s perfect for most retail stores. As someone who worked in retail for a long time, I’ve got the scars to prove it. Christmas music still makes me cringe a little bit. We had a server, quote, unquote, which is just like a robust desktop under the counter, and to replace it with this wouldn’t be that much of a big deal, especially for the retail.

[00:26:55.160] – Jan
And you don’t just take that use case you mentioned. Just imagine a world where in all of your retail locations, you’re just running a piece of the AWS cloud now for you to deploy your applications, update your applications. You don’t think about security anymore.

[00:27:09.910] – Ned

[00:27:10.300] – Jan
Because it’s a fully managed service.

[00:27:12.890] – Ned

[00:27:13.280] – Jan
We update it, we manage it, we monitor it. If something goes wrong, we’ll notify you. So I think that’s a real game changer for going that deep into the edge.

[00:27:24.750] – Ethan
Well, you’re worried about the box from a security perspective in the same way you worried about your AWS account.

[00:27:30.420] – Jan
You don’t have to worry about the platform.

[00:27:31.820] – Ethan
You have to worry about everything that’s riding on top of it.

[00:27:33.860] – Jan
And what we did, by the way, with Outpost is obviously Alpos runs outside of our secure facility. So we had to spend a lot of time thinking, how do I make that outpost that runs? I don’t know where it runs. I’m shipping it out of a box somewhere. How do I make it absolutely secure? As secure as if it was running in my region. And what we’ve done with Outpost is we introduced this NSG, which is a key that sits in the outpost. And if you run your application and you want to send it back to me, you want to make sure there’s nothing on that box that’s your data that goes back. And what you do is you pull the NSK key out, which will destroy it. And that is how everything was encrypted through that hardware key, that the minute you pull it out, you destroy the key, and everything on that device becomes unreadable.

[00:28:21.210] – Ethan
There are also supply chain protections on that box, like TPM, this sort of thing.

[00:28:25.970] – Jan
Absolutely. There’s various security technologies that monitors all parts of the inside.

[00:28:31.730] – Ned
Yeah, one thing we didn’t cover is lambda. And I’m curious, is lambda also an option for. Outpost today.

[00:28:38.890] – Jan
You can run lambda functions, whether it’s in the region, local zones, outpost, the exact same way you run it.

[00:28:47.100] – Ned
Okay, but we’re starting to come up on time. I want to see just check in with you and see if there are any other announcements that came out during Reinvent that you’re super excited about or anything coming down in future road map that you’re allowed to talk about.

[00:29:00.080] – Jan
Oh, we talked about that before we started. Right? We spoke about the local zone. That is an exciting piece. We are continuously to see fantastic growth with the AWS P 5G service. That’s our own AWS private 5g service. So just quickly for the listeners, what this is. Do you think 5G, you think telco, think big systems, right? You know, hospitals, airports, I mean, lots of money. What we wanted to do is we want to make it as accessible to everybody. So we took that entire technology and we said, how can we make it a fully managed service? So AWS P 5G, you can go on and you can order the system. We ship to you all the hardware and software needed for you to stand up the 5G network. And so you can stand it up, you can you can configure it, and you can run it. And then secondly, from a business model, we introduced the cloud economics, where you also pay for the consumption versus paying for the system upfront. So this is really exciting, and we are seeing a lot of adoption of people now having access to this technology. And it’s fascinating to see how our new use cases are being.

[00:30:16.450] – Ethan
That was my question is use cases. I’m imagining IoT factory. These are the big ones that pop to mind, is that typically this is.

[00:30:23.220] – Jan
A surprising one for us was events because it’s so easy now to stand up a 5G network. We use them here. By the way, at AWS Reinvent, our Events team uses the AWS Private 5G for the live streaming cameras. They have wearable cameras they walk around with, and they have fixed cameras. But it’s so easy to stand it up, provision it, run it. And what we did with AWS P 5G, we made it fully API driven through the AWS console. So things like your IAM for managing identity, that device now is just another identity in your IAM. You can manage it, you can decide who have access to it, who can it talk to. So it just feels it’s just completely API driven. And you use the exact same tools used today for developing applications to manage your network.

[00:31:15.030] – Ned
All right. Pretty cool. That does sound very cool. If folks want to get started or dig into more in the area that you specialize in, is there a good landing page to send them to?

[00:31:25.960] – Jan
Absolutely. So if you go to AWS, you’ll find a landing page for local zones. They’ll tell you everything about local zones where they are available, what services or support on local zones. You’ll find a landing page for outpost. We’ll take you through the outpost and you also find a landing page for the AWS P 5G service, private 5G service that I just mentioned. So all of that has a landing page that gives you all the details you need.

[00:31:48.230] – Ned
All right, awesome. Well, Jan, thank you so much for being a guest today on Day Two cloud. And hey, listeners out there, virtual high fives to you for tuning in. If you have suggestions for future shows, you can hit either of us up on Twitter at Day Two cloud show. We both monitor that account, or you can fill out the form on our fancy website, day Two Cloud IO. Thanks again to our guests for joining us. And just remember, cloud is what happens while it is making other plans.

More from this show

Episode 180