 Hello. My name is Roshan Gurupati. I'm from OneConvergence. So here to talk about basically network and service overlays for SDN. So there were a few other stocks earlier, so that kind of laid the context. So that makes my job easier. So hopefully that gives you an idea of what network overlays are all about. So essentially, so we do network virtualization based on an overlay model, the IETF-NVO3, similar to some of the other vendors that presented here today, but we go beyond that. So we beyond the network overlays with this concept of service overlays. So essentially what we're doing is on top of the network overlay for layer four through seven services, we're creating this service overlay. So where we're doing, that is where we're bringing in our value add, our differentiation. So we call this the service delivery platform. As part of this, the kinds of things that we do are basically insertion of a service, you know, L2, L3, chaining of these services, basically. So when I say service overlays and services, I'm actually referring them to network services. So network services like VPN, firewall, load balancer, WAF, IPS, IDS, these types of services. So that is what we're focused on. In addition to insertion and chaining of these services, so we're doing things like auto scaling with elastic load balancing for this service layer. So I actually have a demo that will show you through a live demo that is running right now of elastic WAF service. So, and also we offer distributed services like distributed routing and so on. And we're also working with some vendors like our partners like Cavium Networks, FreeScale and others in terms of even acceleration in terms of scale up in addition to scale out on some of the services API functions. So basically we have a concept of a template or a policy model to drive this whole network and service overlays. So we are also co-contributors in the group policy working group at OpenStack and as well as Open Daylight working with some of the other vendors in coming up with a policy model. So essentially you can have a template that describes the topology, the services and the policies and we drive that to realize the network overlay as well as the service overlay. So this is pretty much our solution. So if you look at it here at the bottom, if you have a shared, basically interconnected IP switches or layer three switches and V-switches in the servers, you know, like an open V-switch for instance. So these together the interconnected IP switches and the V-switches together form the shared infrastructure. So basically on top of this, we are creating this virtual tenant domains with our network and service overlay controllers. So we have our own controller, the whole platform is a software-only platform. So that can be deployed on any hardware infrastructure. So basically, so within this, so we're able to create these virtual tenant domains and then based on the topology, the services and the policies that are basically described in a template, we can automatically instantiate those arbitrary topologies and services and policies into each of these tenant domains. So basically, each domain is fully isolated from each other. So for instance, tenant A could have like a three tier network, tenant B could have some test and dev type of thing, whatever you want. So this could be for an enterprise, it could be different departments, it could be finance, it could be HR. So a completely isolated domain and you have networks basically services and policies based on their needs. So and there, as I mentioned in my earlier slide, so we are bringing in our value add on top of this. We believe our differentiation with respect to other vendors in the service layer. So layer four through seven. So things like auto scale for instance. So we have our own, this whole thing is a software only platform. It is integrated into OpenStack Neutron. So we provide the Neutron backend to realize, to do this network and service virtualization. So as well as we provide orchestration in the front end. So I'll show you what the UI looks like. So basically, we have our own network and service overlay controllers as I mentioned. They're fully distributed, clustered controller, supporting HA basically and clustering. So some of the other things is we believe in open architecture and leverage standards where they exist. So we're using basically OpenStack. We're using OpenVSwitch, OpenFlow, OVSDB, VXLAN. So in a way, we're agnostic to the tunneling mechanism. You can use VXLAN, NVGRE. So basically, we're using all open source stuff but really adding our value and innovation in filling the gaps and where we need to. And likewise, in some of the areas where things are missing, we're working with OpenStack and ODL in helping fill some of those gaps. So a couple of use cases that we're working on. Again, this product is not, we have not fully announced the product but what we're working with some early customers. So one of the customer use cases that we're working on is basically offering security as a service from the cloud as opposed to premise offering it in the cloud. So essentially for this customer scenario, what they have is they have a multi-tenant platform where they're using the shared infrastructure. So creating customized security chains for each of their customer based on what they're subscribing to. And with our platform, basically, elastically scaling this security service or service chains up and down, leveraging the same infrastructure. That is one use case that we're working on. And another use case is this infrastructure as a service. So basically it's a data center operator. So wherein they provide some of their network services and along with that they offer supporting them with a model of bring your own service so that you can basically insert into one of your virtual tenant networks basically. So the carrier NFE use case is pretty much another use case for this so that we are involved with. So now I'll kind of show a little bit of the UI. So if you look at this, this is the default OpenStack horizon on which we added our knock admin tab. So and if I go in there so you can see so basically, so right now I'm logged in as a knock administrator. So which means I have visibility across all the tenants. If I log in as a tenant admin, you'll only see things specific to that one tenant. So these are some of the things that are showing here. And basically you can look through the overlay networks, the services, we have logging, reporting, debug tools and various things. So we have a traffic pattern. So one of the things that we're showing here in our booth and I can just pull up some of the screens is this topology that is described in a template. This has a web application firewall that is based on Nginx and pretty simple topology that we have described it in a template and we instantiated it in one of these virtual tenant domains and we have a traffic generator that is running so with a varying pattern and based on and this particular template has autoscale policies attached to the web application firewall. So based on this autoscale policy we're scaling the web application firewall instances, essentially horizontally scaling, spinning up more instances and modifying the network to do that and then basically elastically doing this up and down. So I'll show you, so this is the traffic pattern that is running. As you can see, this is fluctuating in this manner and based on this, you can see that we're autoscaling the instances, this is showing the number of instances. So they're going up and down. So I have other snapshots that I took yesterday so which has a little better pattern that is running. So if you look at it here, this is the traffic pattern that is running and this at the top that is showing the packets per second at the bottom we're showing the number of instances of the service. As you can see, so this here, this corresponds to this and then this corresponds to this and then all these spikes here. So you can basically configure the autoscale policy. There is a cool down period and things like that set up and right now we support about four or five parameters that we can scale things on, connections per second and bandwidth and a few other things and we're going to add more and more things. So these are some of the things that we're doing basically in terms of adding value beyond what some of these others have talked about in terms of providing a basic network overlays. So we are focused and innovation is while providing the network overlays, so really adding value in the service layer. That is what one convergence solution is all about. And as part of this, we're supporting open source services whether it is HAProxy, NGNX, PF Sense, things of that sort, as well as we're working with OEM services for load balancer, firewall and WAF. So we will be, we're working with a bunch of these partners. So likewise, our solution will be we can work with Red Hat and Canonicals, OpenStack distros. So essentially we can basically install on top. So that's pretty much what I had and there's a lot more to the UI in terms of orchestration and other things that we provide. We'd be happy to show you a demo at our booth or if you can contact us, we could do a WebEx or demo things and we can even provide a free evaluation license for trying these things out. So with that, I pretty much conclude. Thank you.