 Let's go. All right, hey, thank you everyone for joining. My name is Harshal Pimpalkhote. I'm a product marketing manager at AT&T Networks. And the topic for today's discussion is dynamic L427 services for OpenStack Cloud data centers. So we are roughly 15 minutes. What we're going to talk about is some of the challenges that our current data center services face in terms of legacy approaches. We're going to take that as a basis and then talk about what are some of the approaches that we can take as we move towards a cloud data center. This includes dynamic service provisioning, includes automation, integration with orchestration. And we're going to close it out with the AT&T Networks OpenStack integration. So that's briefly the agenda. I think we have 15 minutes and we may have some time left over for questions. We have our entire team over here so we can take questions as they come up at the end of the session. So with that said, before I jump into the topic of discussion, oh, what is AT&T Networks? What do we do? What are our products? So AT&T Networks provides products and solutions so that allow customers to accelerate, secure, and optimize their network applications. And this is done through our products which are in the realm of ADC, CGN, and TPS. So the ADC provides, or the application delivery controller provides acceleration, performance, and security. The TPS product line provides the threat protection systems or the perimeter security and the CGN provides service provider with the CGN capability. So that's roughly the product line. All of these products run on the ACOS platform. That is, in my mind, that is one of the key innovations that AT&T Networks brings to the market. And we're going to talk about it in a couple of slides as we cover the discussion. But ACOS platform with the shared memory architecture and the flexible traffic accelerator provides efficient utilization of multi-core CPU architectures. So that's something that we would like to talk about today. And finally, with these products and solutions, what we are able to do is provide support for multiple IT delivery models. So if you're looking at dedicated networks in case of private clouds, we have high performance appliances that can work as a virtual chassis to provide a single solution. We have solutions for managed hosting environments where you can leverage some of our multi-tenancy characteristics, the application delivery partitions to come up with a managed hosting solution. And finally, the topic for discussion today is cloud IAS. So we have a cloud services architecture that we recently launched in January that can be used for cloud infrastructure as a service offerings. So that's the brief background on AT&T Networks. And here's a quick company snapshot, if you may. AT&T Networks was recently listed on NYSE. So we went IP at roughly two months ago. And it has been a phenomenal time. As you can see, our customer growth has tripled in the past three years. Our revenues have almost tripled in the last four years. So it's been a great ride, very interesting times. And we hope to deliver even more innovation as a public company. So that's AT&T Networks for you and the products. Moving along, or rather jumping into the topic, what are the challenges facing the industry today in terms of L4, L7 services? So if you briefly take a look at it, the L4, L7 services that exist today are static, they're inflexible, and they have to be manually provisioned. So they're inflexible by design, and since they have to be manually provisioned, it takes a significant overhead. The change management can be days or even weeks before you can get a single change implemented in your L4, L7 services. So here's a quick example that I'm sharing with you. What you have over here is a server load balancer running some QS parameters, a WAF, and a DDoS solution on top of that. If you wanted to take out even a single element from this chain and move it elsewhere in your data center, like I said, it could take days or weeks. This is a fundamental problem, a fundamental challenge that hinders movement towards cloud. What you really want in a cloud is a dynamic service provisioning. And that's one of the key areas that we're going to touch upon today through dynamic service chaining. You also need automation and scalability, and you need automation for a quick reaction or agile environment. You need scale as you leverage economies of scale. And the goal here is to turn around things quickly for the administrator. So that's the premise, that's what we're approaching this from. And if you contrast this with what's required, so you have isolated services on one side, like I described to you earlier, you have silos of services, you have monoliths of services that cannot be easily tweaked or easily changed. As we move towards the cloud, dynamic services is taking over, and this manifests itself in a resource pool of services. For those of you who might be aware, one of the five key attributes of the cloud is resource pooling. That's how you achieve a very efficient utilization of your resources, whether it's services, whether it's infrastructure, whether it's people. So like with everything else, with services we are trying to resource pool them into a single pool, so that these services now can be consumed by the applications as they require. Part of the application, part of the service, entire service, or maybe they want to spin it up or spin it down. So that is the dynamic aspect of services that we are referring to here when we talk about dynamic L4 to L7 services. So how does this take shape? This takes shape through policy-driven infrastructure. What we envision is the business needs will be converted to a policy infrastructure. They'll be converted to rules or policies that will be directly fed into the private or public cloud infrastructure. And from there on, you get a simplified demand, simplified on-demand consumption model so that your applications can self-service themselves or the IT administrators can automate this to the next level. Also, we are looking at a heavily multi-tenant architecture. So even with private clouds, as lines of businesses go into behavior separate tenants, you are looking at a very heavily multi-tenant architecture. And finally, to leverage the economies of scale associated with cloud, we are providing with scale-out architecture. So that's how we envision, we briefly talked about the problem, and this is how we envision things changing as innovations take place and we move toward a more dynamic L4 to L7 services. All right, so just wanted to share with you this quick animation. What you see over here is multiple tenants, multiple cloud tenants show up on your network and your orchestration platform is provisioning the compute, the layer-to-layer three networking and the storage piece of it, right? The cloud orchestration platform may communicate with any SDN controller and push down aspects such as VLAN, gateways or QS parameters to the layer-to-layer three network, right? To the SDN network fabric. What's missing in this picture is the dynamic L4 to L7 service chaining. Organizations today need a way to automate and scale. There are L4 to L7 services and leverage the same benefits that they're getting from the orchestration platforms for their layer-to-layer three networks, for their compute, for their storage. So that's the gap that ATEN networks is trying to fill over here with our vision and with our products. All right, so I briefly touched upon dynamic service chaining. So what is dynamic service chaining? It has been, I was watching a couple of other presentations right before me and it has been a term that has been used more recently in the past. So dynamic service chaining is a broad term and it can apply in many areas and manifest in different forms. So for us, dynamic service chaining allows our customers to spin up or spin down services, to consolidate services, to share services across different applications and also to automatically provision them when a new application comes along. So that is the basic scope of dynamic service chaining as it relates to us. So what you see over here is a multi-tenant architecture with Coke and Pepsi and other tenants. And Coke is using security as the service, Pepsi is using IPv6 as a service for its applications. And what our premise is, these services should be able to permeate across boundaries, across boundaries of governance, across boundaries of control, across boundaries of networks. So as the application moves, you have a truly virtualized and a mobile infrastructure. And the underlying operating system, in this case, the application gateway operating system, should be able to support this kind of mobility, this kind of portability. And then finally, the platform, it should be supported on the physical appliances, on the virtual appliances, and also on the industry standard hypervisors. So that's a view of dynamic service chaining as we see it. So just to summarize it, what are the design requirements for L4 to L7 services? I roughly categorized these into three buckets, this agility, the scale, and this reduced total cost of ownership. Agility today is table stakes. For those of you who may have attended yesterday's keynote speech, customers used to look at fast, good, and cheap. What they're looking for now is fast, fast, and fast. If you give them fast, they can convert it into cheap, they can convert it into good. What they're looking for is an agile infrastructure, an agile policy-driven infrastructure. So agile service delivery is one of the key design concepts that we have at ATOM Networks that provides for simplified consumption model, programmability, and we talked about orchestration and SDN integration, but that's just one aspect. As applications move across governance domains, and as services move across governance domains, you need a portability mechanism, which is API-driven, so that you can get that kind of agility across different private, as well as public, cloud data centers. The other is scale. One of the fundamental reasons we are going to the cloud to a public or private cloud is to leverage the economies of scale, and application delivery at scale is a key item. You have to linearly scale with performance, and I'm going to talk in a bit about how the A-Cloud platform allows you to do that. Also, you need consistency of services, so you should be able to enforce SLAs for your security, for your compliance, as time passes, and you need reduced total cost of ownership, so you need meter consumption, as well as simplified management. So the A-Cloud services architecture, we took all of these design precepts to the table and created the A-Cloud services architecture. This is something that we launched earlier this year. It's a portfolio of products and solutions that provides our customers with the cloud offerings, the L4127 services offerings, necessary in a private or a public cloud infrastructure. And these are based on our hardware and virtual appliances. They are based on a pay-as-you-go model to reduce the total cost of ownership, and we have integrated, or we have integrations with SDN and orchestration platforms, including OpenStack. So I'm going to talk about the OpenStack integration in a bit. So in interest of time, I'm just going to quickly refer to this. So with the appliances, you can have different IT delivery models. So you can have a dedicated data center, or you could have a managed hosting, or a cloud infrastructure as a service. And this is the one that we are referring to over here. The v-thunder appliances or the virtual appliances allow that kind of mobility and the scale and agility necessary for a cloud infrastructure as a service offering, along with the pay-as-you-go licensing. So moving forward, one of the key innovations that I described to you was the ACOS platform from ATEN Networks. There are really two components over here. One is the flexible traffic accelerator that evenly distributes traffic, evenly distributes flows across all the CPUs. As we move to a multi-CPU architecture, it is critically important that we utilize all of the cores, all of the CPUs evenly. And the ATEN flexible traffic accelerator is the key component in the ACOS operating system that does this. And the second piece is the shared memory architecture. So in case of ATEN products, in case of ACOS, we have a single shared memory architecture. So we eliminate the need for any IPC communication, which can be process-intensive. And you have a consistent view of the memory so that you can apply the policies quickly. You can do so in an agile manner. So that's the ACOS platform that we have powering all of our appliances and all of our ACloud offerings. As far as the licensing models are concerned, we have both a perpetual license that you can buy once and use it forever. Or you could have a monthly license, more like your cell phone plans, or a utility license, just like your utility bills or your electricity bills. So you have a RBM and a UBM. And this forms the basis for your pay as you go in the cloud infrastructure environment. We are also available on the AWS Marketplace. So you could deploy a v-thunder on the AWS Marketplace as well. So those are the billing options. And these are my last two slides. As far as OpenStack integration is concerned, what we have done is we have used the Elbas Neutron extension that you see over here and that Elbas Advanced Service Plugin to insert the A10 Elbas driver. And by doing so, now you can leverage all the different, or you can use all the different appliances that I mentioned, the v-thunder, the hybrid, and the physical appliances in an OpenStack environment. And the goal again over here is to provide the same functionality, the same dynamic L4 to L7 services in an OpenStack environment. So, how does this all look when you have the OpenStack piece, when you have the SDN piece, when you have the virtual as well as the physical appliances? I'm just giving you an example over here. So you have multiple tenants. You could use your physical appliances, get into the SDN fabric. You have a v-thunder sitting on a v-switch and you get the server load balancing functionality over there and you move from there to the WAF or to the v-switch again where you get the WAF functionality. So you have three different ADCs in your service chain. That's one aspect of it. If you look at the other service chain, you have a hybrid appliance sitting on your left providing you with the server load balancing capabilities again as the traffic goes into the SDN fabric. So that's the overall view as we see it from an ATIN perspective. If you have any questions, please let us know. We are in the booth right there in booth number 313. We also have an OpenStack demo that's going on. So we'd be more than happy to answer any of your questions. Thank you for your time.