Hybrid cloud storage and edge compute

All right. Hey everyone. Good morning and welcome to day one of what's gonna be a super action packed re 2023.

We're excited that you're all here joining us today in our session on hybrid edge storage and compute.

In this session, you're gonna hear from us about all the innovations and all the cool things that we've been working about working on in AWS hybrid edge over the last year or so, including pushing our AWS edge boundaries all the way to the space. Yes, you heard that right. But we'll get to that in just a bit.

My name is Linda B. I am a senior product manager on the AWS Snow Family and today I'm joined by my colleague, Mark Nan, principal solutions architect on AWS Snow. And I am super thrilled to introduce our guest speaker, Ted Suzy, chief technologist with Wat in today's session.

I'll be talking about the AWS cloud to edge continuum. Why and how it's important and how AWS has essentially built on this continuum. I'll be talking about two AWS services, AWS Outposts and AWS Snow Family. And I'll be using these two services as an example to show how we have extended AWS infrastructure and services all the way to the edge.

Finally, I will invite Ted to come in and talk about how WireSat is doing all the innovations, how they're doing some of their innovations on the edge using AWS products.

Finally, as always, we are happy to take any questions at the end and feel free to meet us after the session as well.

So as with all of Amazon, everything at AWS begins with our customers, everything we do begins with our customers and everything we do is for our customers and on the hybrid edge space.

One of the things that we've been doing is very intently, been listening to our hybrid edge customers like you over the last few years, we've heard three main requirements from our customers.

The first one being the need for AWS to extend its services and infrastructure from AWS region all the way to the edge. And this requirement primary stems from customers use cases that require them to run low latency applications on the edge. It also includes use cases that involve migration and modernization of their workloads from edge to the cloud as well as addresses some of those use cases around local data processing as well as data residency use cases.

Our customers like you have told us that with the existing solutions that they have available in the market today, they face several challenges more specifically, they find it complex to set up the infrastructure using multiple solutions. They find it cost intensive and they find it challenging to maintain their applications and solutions on prem and at the edge.

So we double clicked into that a little bit more. And what our customers essentially told us is that they want consistency between the experience that they have in AWS region to on prem to the disconnected or to the far edge.

More specifically, our customers wanted consistency in five different areas. The first area where our customers are seeking consistency is in needing the same, reliable, secure and high performance infrastructure that is used in data centers today.

The second thing that our customers like you needed was the same ownership model and operational consistency as the cloud. Our customers like you also requested or asked for API tools, automation and deployment consistency across region as well as the edge.

And finally, they asked us for the same speed of innovation at the edge as available in the AWS region today.

So we've been working with our hybrid edge customers for the last several years. We've been hard at work at extending AWS cloud with new services and capabilities that bring applications closer to where you are running your applications today to where your equipment is available to where your customers are and to locations where you are collecting and processing your data.

Today, it starts with the AWS regions and the AWS global infrastructure. And then we extend that to metro areas and telco networks using services like AWS Local Zones and Wavelengths moving on to on prem with AWS Outposts.

We offer solutions from data centers to remote and many times at the far edge at rugged locations using AWS Snow Family with all of these services available. We have extended AWS's capabilities from in region to on premises and the far edge.

And if you notice at the top with the arrow that's there, you'll see that it extends both ways. It is now the case is that because we realize that sometimes customers workloads are born in AWS region with the need to extend the same workloads and applications to the edge and sometimes applications are born at the edge with the need for customers to connect to AWS regions based on when their needs arise.

Now we'll double click a little bit more into two specific areas within the AWS hybrid spectrum or the cloud to edge continuum. We'll be looking at AWS Outposts and the on premises and how we are bringing AWS infrastructure and services on premises.

And we'll be looking at AWS Snow, which is where we're bringing AWS infrastructure and services to the far edge.

AWS Outposts enables you to develop your application once and deploy it in your data centers and in the cloud, we AWS Outposts enables you to do so because AWS has extended the infrastructure and services and the same technology that we are using within our AWS data centers, which is AWS Nitro Systems by making those same hardware and software available within your on prem locations.

This enables you to use a fully managed AWS solution that is monitored and operated by AWS as if to provide the same experience as the AWS regions. And it also provides a single pane of management in the cloud using the same APIs and tooling that you're using in AWS regions today.

Within Outposts, which is essentially a family of AWS managed devices. We provide two configurations starting from the one U and two U Outpost rack Outpost servers all the way up to the 42 U Outpost racks.

While the Outpost racks are suitable more for use cases where you need to run your applications with lower latency, close to your on prem locations or your co-located data centers to run maybe applications like financial analysis gaming or maybe even like online casino systems.

The Outpost server on the other hand, in its one U and two U form factor is more suitable for use cases where use cases where you have a space constraints, right? So location, think of locations like point of sale systems in retail spaces where you don't want like a large rack sitting in your systems, you can use the smaller Outpost servers in the one U or the two U form factor.

Over the last few years, we've been working to expand the number to expand the number of countries and territories where AWS Outpost is available today. Outpost is available in over 76 countries and territories across North America Europe and Asia Pacific. And we continue to extend the services and capabilities that are available in Outpost today.

If you look at the slide, it kind of gives you an overview of what are the services that are available on Outpost today. You can essentially run a variety of these services on Outposts to meet your low latency and data residency needs while connected to regions for all AWS services.

For example, you can use S3 and Outposts for object storage. You can use EKS and ECS on Outpost for your container use cases. You can also use RDS and Elastic Cache on Outpost for highly available and highly performing databases.

And we are constantly innovating and launching new capabilities on Outposts to better support our customers needs at the edge.

Now, let's talk about how we are extending AWS cloud services to the furthest and the smallest edge. But before we go into that, let's talk about what is the rugged mobile edge?

Our customers like you have told us that there are many edge locations where they're running their applications. But these edge locations have limited to even sometimes no connectivity at all to the cloud. These locations include construction sites, oil rigs, fleets of ships, autonomous vehicles airplanes and even rapidly deployed military forces.

And we've taken AWS infrastructure to, like I said, the most extreme environments, which is the space and we'll talk about that in just a bit.

Looking at the AWS Snow Family, which is essentially a family of hardware devices that are portable and rugged and are meant for customers to deploy their AWS based solutions to transform, capture and move data from the rugged mobile locations back to the cloud.

Within the AWS Snow Family. We have two form factors. The first one is the small tissue box sized AWS Snowcone which is a compact and purpose built device. It's portable and it's used for it's used outside of of your traditional data center use cases, Snowcones are meant for use cases such as satellite imagery, aerospace use cases.

And then the second form factor that we have available is your larger carry on luggage size. AWS Snowball Snowball device is available in two form factors. The first one is the compute optimized Snowball Edge, which is more suitable for use cases where you have compute intensive workloads running at the edge such as AI/ML, predictive analytics and so on and so forth.

And then we have the storage optimized device which is essentially better suited for use cases where you're collecting large amounts of data at the edge with the need to process the data locally and transfer this data back to AWS for storage retrieval and other use cases.

And both of these devices like I mentioned are extremely suitable for, are suitable for use in extremely extreme environments and conditions. And these devices can also be clustered and scale and connected to provide additional capabilities to our customers.

Now, I'd like to invite Mark to talk about some of our customer use cases with AWS Snow devices.

Thank you, Anita. Good morning, everybody. So the first I want to provide two examples of use cases. The first example is a, a tremendous human impact example.

I remember back in high school, that was a long time ago. My ninth grade civics class, the teacher asked the class what defines a civilization. And it's not that they have religion, it's, it's not that they have traditions or they live in a society together. But the definition of, of what makes up a civilization is that they have a written language.

Therefore having that written history where they can pass along the generational knowledge from generation to generation. That's that learning and build upon that when you think about what that translates to today? That's, that's data, right? That's information.

And, and so imagine if you were to, to lose that information, imagine if, if the government lost your social security number or imagine if you couldn't collect your monthly pension check because you didn't exist.

So you know that that's very important. Last year Ukraine came to AWS under let's call it extraordinary circumstances. And they asked us to help them preserve their data. That's on premise into a more secure, more durable location, help them so that they can access that data and build upon that later on in case something happens, right?

In case, you know, Kiev gets overrun or something like that, you know, we want to preserve that data. So this is a perfect example of where Snowball Edge shines. The simple edge that that Anita mentioned a storage optimized device. It really shines in the circumstance, but also it was the the volunteers on the AWS disaster response team and the solutions architects, they shine there too.

They brought many Snowball Edge storage optimized devices to Ukraine. They work with the Ukraine government to copy their data onto these devices securely and safely transport these devices across the border into a location where they can be sent to an AWS region and get ingested into the region being safely and securely in a location where they can then access that data and use that data.

It had, you know, a profound effect on their capabilities. So it's a really cool example and it resonated very well among many of the other NATO and European Union countries. And the fact that a lot of them are re-evaluating, you know how they look at their on premise data, making sure that they have copies of it in other locations. And not only that it also affects the way they think about compute, right? They've got critical infrastructure that they, that they use and what if there's a nation state attack? You know how making sure that that critical infrastructure still is able to provide services for their citizens. So making sure they're being able to run disconnect for certain periods of time. As an example.

The next example is a is a fun example and this is kind of near and dear to my heart because I was directly involved in this project. The International Space Station was first launched. The first components of it was it was launched in 1998. And over the years, it was multiple countries contributed to it and added different models to it. And it grew um over the course of time now it's, it's reaching the latter days of it. It's scheduled to be decommissioned by year 2030 and it's going to be replaced by private um space stations, private commercial space stations.

And one of those such companies is Axiom Space. Um Axiom Space recognized that the cloud has a profound effect on earth, right? Cloud computing uh has enabled a lot of companies that were born in the cloud that wouldn't, would not have existed the way they do it without that, that cloud connectivity. Um so how do we take some of that uh cloud capability and bring it into space? Um Axiom asked us, hey, work with us, either of us we want to set an example of just, you know, a notional experiment initially in the early phases of our development. How do we bring that capability to our project in, in our space station?

Um they looked for a device um that uh was already really available. Um something that uh you know, it takes years to develop uh devices that are suitable for space. So what's already available, what's off the shelf? Um but also they wanted, they liked how AWS uh had uh had the cloud already, but they wanted something that uh could, they could leverage, customers could leverage the same API the same tooling to be used in, in the, in the space, right?

Um so we selected the the Snowcone device and um the Snowcone device is something that they as rare revealed for everybody. Um they ordered it off the region, it was delivered to their facilities. Um and we went through a number of tests. Um I spent multiple weeks at uh in Houston at Johnson Space Center. I ran through things like um penetration testing, vulnerability testing. Um but other tests as well, things that you may not consider on earth. Uh because we don't, we just don't think about, for example, uh did you know that hard drives come with a built in accelerometer? And uh it's, it's there because when it detects a fall, it uh it's it's gonna lock that drive heads to make sure that, you know, if there's an impact it doesn't, uh, corrupt the data, it doesn't damage the, the drive platters well, in space at zero gravity, guess what? You're always falling. Right. And so the, the drives don't work. And so you have to either use specialized firmware for the hard drive or you can use SSDs. Right. But SSDs have their own issues with space and they're more susceptible to, uh, cosmic radiation, solar radiation.

Um other things like did you know that um uh noise pollution is a problem in the space station? You know, you can't have, you can't have devices that are allowed there. So we had to go through acoustic test. So we make sure that, you know, an astronaut is not uh as different appliances and devices in the space station, all aggregates noise, you know, you don't wanna, you don't wanna contribute too much to that.

Um yeah, so a lot of lessons learned. Another really cool example is, did you know that uh disconnecting a USB C connector um causes a micro spark and um and can also cause a piece of molten metal to, to come out now on earth. You know, we don't care, it's gonna cool instantaneously and drop to the floor. No one knows, no one cares uh in space. However, that particle is going to potentially float in the capsule. Uh it could go into an experiment. Um it could um it could uh you know, go on someone's eye. So um so what we did was, you know, they select the the Snowcone device um went to multi test all the tests passed after a month.

Um you can see in that image there that is the actual Snowcone device floating majestically in the cabin with a beautiful mother earth in the background. Um I love that picture. It's the, it's, it's not a photoshop or anything that's actually how it looks. Um and as Nita mentioned, it's about the size of a tissue box. Uh we ran some machine learning models on there and uh it was able to look at the images, detect the, the, the objects and report on them. Um it's still in operation today so that uh that Snowcone is on the International Space Station. You can and there are customers to pay Axiom to load a workload on there and run an experiment. So a very cool example, very fun example.

The next uh so here is, as you can see, there's a list of, of services available for Snow. And uh there, there are compute services like EC2 and EKS, there's storage services like EBS S3 Tate Gateway. And there are also IoT services like IoT Greengrass, you can run Lambda on IoT Greengrass, you can run IoT Sitewise Edge.

Um the key thing to remember here is that um you can use the same tool sets and the same um custom code to interact with this. For example, if you wanted to launch an EC2 instance in the cloud, you could do a ec2 run instance. And along with the parameters for that, you can use that exact same syntax for launching an EC2 instance on the Snow device.

Um you just have to change the endpoint, same thing with creating a volume um or storing objects. So um that's the the key benefit of Snow. The other thing to keep in mind is that Snow was designed to have a local control plane. What that means is that you can put a Snow device and connect your laptop directly to that device and interact and manage that device without having any internet connectivity.

Um so now it's operable in the space station, it's operable on a ship um on a forward operating base. Um so great examples of, you know, why Snow has that, that niche play for, for certain use cases.

What I want to do next is I want to highlight two features that were released this year. Uh the first of those features is Amazon S3 compatible storage on Snow. Um for those of you who are familiar with Snow already, you're probably asking, hey, Mark, don't you already have S3 on Snow?

Um yes, we do, but this is this new uh implementation of S3 builds upon that. Um what we're doing here is we're adding additional functionality to make it more similar to what you're experiencing on. Uh on the outpost. So what that means is um you've got more uh storage, you can cluster your devices together, so you can deploy a single device and there is some additional redundancy there using the way we strip the drives. But there's also you can cluster them from three devices to start off with.

Um as you need more capacity, you can scale it out and add more devices in increasing your storage. Uh you can go up to 16 devices per cluster. In addition to adding, you know, scaling and all that's going to give you some more durability, some more reliability. We also allow, we also open the doors to some of the other APIs that that you familiar in the region and on, on, on an outpost and things like you can do get objects, push objects um and some of these other API commands

Um as you know, security is always day zero at Amazon and AWS and so all the data is encrypted, you don't have a choice not to encrypt. So you can pick either SSE S3. So the service side encryption using the AWS managed S3 keys or you can provide your own customer keys and, and use that for the encryption. But you know, we we make sure that, you know, encryption is on uh by default and always on and you can't choose not to and some additional S3 bucket object management things.

For example, before you weren't able to do your life cycle policies. Now you can add life cycle policies to your S3 buckets and making changes and deleting things on in different times.

The next example I share with you that was released this year is Amazon EKS Anywhere on Snow. And the way this works is that you can order a Snowball Edge device and have it preloaded with an EKS Anywhere admin instance. So EKS for those people who don't know is Elastic Kubernetes Service. Um so this is a service that's in the region that allows you to have to manage your Kubernetes clusters. It's also a service on Outpost that's been there for some time. And now this year it's an hour service that's available EKS anywhere on Snow.

It allows you to manage your EKS cluster in a um a fashion where it's not, it deploys things automatically for you. So the way it works is you, you order that device with the instance, you launch that instance and you configure this YAML file, this is your configuration file. It's used to decide, you know the size of your clusters. Uh what type of workers you have, how many providers you have um all that information there and including the networking configuration.

And then you just run a command called uh create cluster. And it's going to go out there, reach out to all the different providers that it has and talks to them deploys the worker nodes builds a cluster, ties it all together. It takes about 20 minutes and and you have the cluster built for you. You can think of it as CloudFormation, the way CloudFormation builds infrastructure in the cloud EKS Anywhere will build and manage your Kubernetes cluster on premise.

Once you have that cluster deployed, you can now build your workloads on top of it. You can customize, customize it as well. We have curated packages for add ons such as um metal lb. If you wanted to add a low balancer, um if you wanted to use your own uh uh worker node image, we also provide instructions for how to do that. If you wanted to, let's say as an example, you wanted to upgrade it, there's an upgrade cluster command and you can upgrade your entire cluster uh very straightforward uh using this method.

Um other things that it automates for you as an example is let's say you have a worker node that dies. Um once that times out between that heartbeat from the ma you know your controller node to that worker node, it's gonna launch a new work load for you. So it's managing that entire cluster for you. Um making it more resilient, more easy for you to, to worry about um taking care of the workloads that run on top of it as opposed to the infrastructure of the cluster itself.

So with that, those are my two examples I'm gonna hand it off to Ted.

Excuse me. Good morning. Those are pretty interesting and cool examples. The snowball in space, I thought that was pretty neat. So I work for Vat's mission connection and cybersecurity group. And part of our objective within our group is to do a secure and resilient networking from the edge to the cloud and everything in between. And so our customers are primarily military and defense. And so part of that is we provide network services. The other part of that is we build products. And so some of the products that we build within our group are our high assurance encryption devices. So these are appliances that allow us to protect data and transit up to top secret. We build free space optics terminals. So think, think of things like laser communication systems, we build five g cellular solutions and we also build what we call our net agility software defined wide area networking solution that's really optimized for the tactical edge. So that's within my mc two group, uh bigger viasat.

So we actually started in 1986 as a small business innovation research awardee, which means since the beginning, we've had kind of our roots in, in dod um since then we've grown up. So now we have about 90 offices around the globe and about $4 billion in revenue. The other thing worth noting on this slide is about 38% of government systems, which is the group that i'm part of the larger group is veteran workforce. And so we take the mission of protecting our soldiers very seriously within bias.

One of the initiatives that we're doing is trying to connect military edge customers to the hyperscale cloud. And you could ask, well, why do they need to be connected to the cloud anyways? And so we actually have um, four reasons for that.

So the first one is what we call situational understanding. And so if I take an example from my own life or your lives, right? We we look at it and we say, you know, your significant other comes to you and says, hey, honey, remember that photo we took on vacation, you know, at the beach two years ago and you say, yeah, of course, i remember um it's on my phone, right? It's on your phone with 10,000 other photos, right? How do you find that photo? And so what i do is i open up my phone, i might search for a person's face, i might search for the location of the beach. I may have tagged the photo with a piece of information, you know, a description that, that i know. And so in the end, those 10,000 photos are reduced to, let's say a handful of photos that myself as a human can look at and say that's the photo i was looking for, right?

It's the same thing for our our military the guys at the edge or the ladies at the edge, they have data, they have lots of data, they actually have too much data. So they have all this information coming at them. What they need is something to help them distill the information. So as a human, they can look at it and they can take action upon it and we call that situational understanding. And so in comes the cloud, right? So the investment that we've done in data analytics and machine learning, those commercial technologies that really make our lives easier and make us be able to process and understand large amounts of data that can apply to their situation as well. So immensely powerful to get them that capability.

The second one is what we call joint operations. So there's a concept called joint all domain command and control. So jd c two and its idea is you're going to connect a sensor to a data producer to a data consumer. And so the data producer could have never talked to the consumer before. So imagine, you know, maybe there's an army sensor and there's a navy ship and they want to talk to each other and they've never done that. How do you connect them together? And that's actually a really hard problem to do in real time. One of the ways you can do that is you can basically bring everybody to the cloud and then disseminate the data from there.

So if we take an example, let's say, you know, we're in a conflict supporting someone else, a hypothetical example and you know, the us is there and we've got this partner nation. And so the location of our troops of our assets we consider secret, right? We hold that near and dear to our heart. That's classified. We're not going to tell anybody who our ships are. It's our information, our partner nation, we consider their location sensitive but unclassified, right. So they're, they're important to us, but we don't hold them with the same regard as we hold our own. Just for whatever reason.

In that case, I've got location of troops and soldiers that are classified and I've got location of troops and soldiers that are unclassified and I've got a soldier and she needs to look at her screen and see everybody. So how do we solve that? Right. One of the ways you can solve that is you could bring those troops from a partner nation into, let's say commercial cloud. And so the their locations aren't commercial cloud and then you could bring the location of the us troops into what we call impact level six cloud or secret cloud, right? This is the classified cloud.

And so from there now you have at least everybody in the cloud, right? So we're all in the cloud. We're not in the same cloud, but we're in the cloud from there. You can do what's called a cross domain where you basically take that information from the unclassified cloud and you move it up in a kind of a guaranteed manner into the classified cloud. And now that soldier, as she looks at her screen, she sees what's called the common operational picture. So she sees all of her forces and she sees all of the partner nation forces in the same location so she can see everybody, right, super important. Everybody wants to see everybody. That's half the half the battle is where are all the people around me. So that's the second reason.

Third reason is to reduce vulnerabilities. Currently, what happens is, you know, they load a set of software onto their compute systems and they deploy with that software, right, configuration managed. This is the version of software gonna use. What happens if there's a vulnerability in that software, right? What happens if there's an exploit in that software that can be taken advantage of where we want to go is we want to go to what's called software factories, right?

So you've got teams that are continuously updating the software, they're monitoring the software and when they understand that there's a vulnerability in the software, they're going to patch the software. And since the software is now cloud native, we can redeploy it as a container, we can push it up to the hyperscale cloud, we can push it to the edge devices and the whole system is patched, right? It's hugely important to reduce that attack surface in real time.

The fourth reason is faster adoption. So the department of defense has looked at the fact that commercial cloud has these technologies and has to realize that there's a lot of innovation going on in the commercial sector that the dod can take advantage of. Right. And so what they've done is they've started to allocate funding and money towards this. And so one of the contracts they have is called jw c. And they've essentially allocated $9 billion to build out unclassified and classified commercial or unclassified and classified processing for dod. And so now we've got cloud resources that are available for government use where they can adopt this type of technology.

Great. Right. So what's the problem? Well, how do you actually get the data to the soldier out at the edge? Um it's a hard problem, right. So these networks are typically intermittent. Sometimes they're being attacked by adversaries. Sometimes the guy setting them up is under duress, right? It's kind of like me standing up here giving a presentation right now, i'm under duress. You know, that, that's kind of how they feel, right? They're, they're stressed out, they're, they're not thinking about everything in all the situation, right? So how do we help them with that?

And so via sat, we're really good at transport pipe, right? That's what we do. So we've got a global sat com network. Um we can get you internet pretty much anywhere in the world. We've got our net agility software defined wiring network router. So we can use our transports or we can use pretty much any transport in the area to get that data back from the edge to the cloud. We have our high assurance encrypts which i mentioned before. So these are really the devices that are used to protect classified data and transit up to top secret, right? So we understand the security piece of it and we have partnerships with people that build the radios, build the compute systems. So we understand the ecosystem so we can kind of build out the tactical network.

Amazon comes to bear with the snowball edge device, right? And so once we have the network and we have the ability to process these workloads using snowball edge, we've created a solution that gives our customers resiliency in their network and gives them flexibility in their mission applications, right? So they have resiliency and they have flexibility in what they want to do and how they want to use the compute to their advantage

So in my mind's eye, if I look at it, this is the picture that I draw, right? So here, here's the vision. And so if you can imagine you're a soldier out there, you may have multiple different types of transporting connectivity. So maybe you have, you know, satellite communications, you've got geosynchronous, you've got low earth orbit um perhaps in the area there's 5G cellular. So there's a high altitude platform which is hosting a 5G radio access network. So I've got 5G coverage and then maybe I have communication through what's called tactical data links, which are essentially the links they use to talk to fighter jets, right?

So I've got this rich network of coms links between the different nodes. And then on top of that, what we're going to do is we're going to layer in cloud native applications. And so we're not going to put these applications in one particular spot. We're going to put them everywhere, right? We're going to build an overlay of cloud native applications and we're going to be able to scale this overlay from the individual soldier to what they call the tactical operations center of the talk. That's the tent with lots of people in it, right?

So one person to lots of people. And so if we look at the picture, what we see here is at the individual soldier, you could do something like AWS Greengrass, right? And that allows you to run small lightweight containerized workloads. And so imagine a soldier, she's out there and maybe she's running what they call a TAX server, which is a server that keeps track of her location and maybe the location of everybody in her platoon, right? She could run that locally. She can host that locally small, it's lightweight, we move up.

So the next step up is the vehicles. So maybe the vehicle we use the Snowball Edge, right? And that's kind of what we're talking about today. That's right now, the sweet spot for mission applications, it has enough horsepower, it has enough compute, it has enough storage to run these instances as, as two or as containers and be able to really get after the applications and services that they use today. So really powerful, really effective.

And then we can go one step further and we can say, well at the TOC, what could we provide them? And then that could be something like the AWS Modular Data Center. And so if you guys are familiar with this, this is probably the thing closest to hyperscale cloud compute in theater, right? This is lots of horsepower, lots of compute cores, lots of storage, right? It's the, it's the huge deployment system.

Um and so now between them, you've given them the ability to kind of scale and deploy the applications where they need to. And so in the end, what matters, right? So the people have the application and the application runs when they need it to, it doesn't matter where it's hosted, they don't care if it's hosted on their person, they don't care if it's hosted back somewhere else. They just know the application works. It's kind of like your cell phone, right? You take your cell phone out, you punch in a phone number, you hit dial, you don't care what 5G base station or 4G based station you connect it to, you don't care what r parameters your phone is using, it doesn't matter to you, right? What you care about is the fact that your call went through.

So this is what matters to our customers. They don't care about necessarily where the data is. They just know that at a certain point in time, they need to access that data and when they need to access it, they want it now. And so that's what this solution is able to build out. And so all the the complexities of the networking and the cloud native applications and the cloud instances are kind of handled through the networking system and through the Snow Edge devices that allow the customers to be able to use the applications anywhere on the, on the battle space.

So what's the solution? So we kind of touched on it, but what we're trying to do really here is we're bringing resilient software to the edge, to those those soldiers as individuals whose, whose lives depend on it. right. And so what we've done is we've taken the Snowball Edges and we've connected them through our, we call our Net Agility SDWAN solution. So that's the ability for us to back haul that data. And then on top of that, we've layered on the appropriate either virtual private networks, if it's commercial or high assurance encryption devices, if it's classified communications and so the result is that the applications can communicate across the network.

So when the network's up and everything's working the way it's supposed to, every individual has up to date information, they have the latest and greatest set of data in all domains, right? Everything is the most accurate copy across the network when the network gets degraded. Right? It's, it's an adversary attacks it or, you know, somebody kicks out a cable right, somewhere in that spectrum, it could be anything. The missions still persist, the applications still work, right? They still, they still run, they still, they do what they need to do. They may not have the most accurate data. It may not be up to the minute, but it's good enough to keep them doing what they need to, right? It's good enough for them to persist in running their applications, running their mission, executing what they were sent to do and that's what our customers care about and that's what's huge for them and that's what this solution has enabled.

So it's been very powerful, it's very effective. It's allowed them to continue to operate and sort of be unaware of the fact that maybe there's degradation in the system. So you got to have an architectural diagram, right? So this is what it looks like.

Um so if I look at this right at the left, we've got our two Snowball Edge locations. And so those are hosting the customer's two workloads we've revised that Net Agility router, it can deploy stand alone or you can put it as an AMI can run inside. And those are connected back through multiple different types of transports. So sat com or cellular, we basically connect those back and those are landing in commercial cloud first, right? And if it was a commercial Snowball Edge instance, you could basically run your workloads there. But in our case, we're going to use GovCloud. So we actually hop into commercial cloud transit through commercial clouds. We're using it kind of like transport, pop out in the US and then pop in the GovCloud. And then from inside of GovCloud, basically, we access the the VPC of our customer and then they can run their EC2 instances.

So now from the edge to the hyperscale cloud, they've got two instances that are secured over VPN tunnels and they're federating data between the two of them. And so this is really huge for them, right? It provides them the ability to execute their mission as we said before.

Ok. So we were able to demonstrate to our customer that this works. Um it provides the solution they're looking for. So what's next? Right. Well, if we look at this, that vision slide, we showed a little bit ago, right? It had Snowball Edge devices, but it also had kind of scaling up and scaling down. So if we look at scaling up, we're looking at the Modular Data Center and kind of how that would look in our customers, what they call concept of operations or CONOPS, right? They're, they're use cases. So what does that look like to them? You know, what does it look like to have that much compute and horsepower in theater to actually execute missions? And so we're kind of working through that the other way is to go, you know, small lightweight.

So we talked about sort of that individual soldier, she had running a server on her person, right? And so we're looking at AWS Greengrass for that. And so our team has been working with the AWS team to basically build our Net Agility router that runs AWS Greengrass on top of it that then runs the IoT workload on top of that. And what's really nice about that is now you've got a package and if you've got a size, weight and power constrained or SWAP constrained platform or they have an existing platform, you can put that in there and you get the whole solution, you get the routing to the applications kind of in one package. And so we're looking at that, we're exploring that with customers as well.

So from my perspective, you know, what is valuable about this is we were able to take a problem and we were able to apply our expertise to it. So Via that's expertise in networking, we were able to collaborate with the Amazon team and enable our customer to recognize their vision, right, and enable them to really focus on their mission, right? Execute their job to be done, do what's important to them and handle the complexities of the systems to, you know, folks like you in the audience and, and us that, that help, help support these war fighters.

I'm gonna hand it off to Nita to wrap us up with the key takeaways today. Thank you.

Thank you, Don. Alright. So you guys have had an opportunity to hear about all the things that we're innovating on the AWS hybrid edge over the past year and we continue to do so. But I want you to take away these three things from today's session.

The first one is that AWS has created a truly consistent cloud experience from all the way from global regions to rugged edge that allows our customers to accelerate their innovation with access to the broadest and deepest set of cloud capabilities that are supported by AWS and our partners.

Second is the increase in productivity and the agility that our customers experience by using the same consistent experience all the way from the cloud to the edge.

And lastly, the large partner ecosystem that we have to enable you to achieve your edge computing or edge storage goals while you're here a three and went this week, which is again, gonna be a super exciting week.

Uh be sure to attend our additional hybrid edge and uh storage sessions. We have multiple breakout sessions, chalk out chalk talks as well as workshops lined throughout the week. Also, we have two really cool demos this year.

Uh so while you are at the AWS village in venetian, feel free to stop by to check out what we are calling the drone inspector. This will enable you can, you can get your hands on with trying out generative air at the edge and using natural language to give commands to the drone in the real world. These use cases, uh the type of use cases is help is used for wildlife or sorry wildfire prevention as well as for disaster recovery and things like that.

And if you're so good at giving your prompts to the gen a model that we have running over on this drone, you also stand a chance to win an AWS branded fleece. So do uh do give it a try.

The second memo that we have is from our AWS disaster response and recovery team whose mission is essentially to support communities throughout the disaster life cycle by improving access to AWS technologies and also bringing it all the way to the edge in the demo that we have available in AWS village. You'll see our rolling lab and how we are rebuilding connectivity and utilizing the technology that we have for disaster mapping and to build solutions for good.

I'm sure it's gonna be an exciting week. Uh it's also gonna be very overwhelming. Uh so you can leverage these additional resources to continue your aw storage and its compute journey. You can take sessions or sorry trainings that are available as well as earn your a w storage badges.

Finally, we have a couple of resources that you can take away to learn more about AWS Outpost and AWS. No family, the two services that we talked about today.

Thank you all for your time and I hope you all have an exciting rest of your week. Thank you.