TMCnews Featured Article
Cisco's Enhanced XR-12000 Router and Multi-Service Blade Brings Virtualization to Service Providers
By Richard Grigonis, Executive Editor, IP Communications Group
Ian MacDonald, director of marketing for Cisco’s Core Routing Business Unit, said the enhancements all relate to the Cisco IP Next-Generation Network.
“You’ve heard us talk quite a bit about Cisco’s IP NGN vision, and now I’m going to tell you about how we’re enhancing the IP NGN to really focus on our ‘Connected Life’ program, this time for working users,” MacDonald said during a recent interview. “We’ll focus on business services and how we’re evolving the Cisco IP NGN architecture to provide a better experience for business services provider customers.”
“The Cisco IP NGN architecture basically has three layers: The Network Layer on the bottom, the Service Layer in the middle, and the Application Layer on top,” MacDonald said. “I’ll focus now on some elements that exist in the Service Layer and how we’re bringing them down into the Network Layer and creating more intelligence for predictable business services.”
“There’s a big business opportunity for service providers to go after the managed services market,” he said. “When we say ‘managed services’ we mean those for small and medium-sized businesses [SMBs] or enterprise customers, primarily branch offices, where they outsource the management of the IP infrastructure, whether its management of the routers or other related equipment. If you scrutinize this opportunity, you can tell that certainly SMBs would like to see that outsourced because there’s a cost associated with it, there’s complexity in the management of the equipment, there’s certain expertise required, and the whole thing is just not part of an SMB’s core competency. These same kind of opportunities exist within the enterprise, not so much within the corporate offices, but more in the branch offices and remote sites of the large enterprises which could be outsourced to a local provider.”
“So there’s a great opportunity for SPs to deliver that,” MacDonald said. “But what they need to deliver to the enterprise customer is becoming increasingly sophisticated. The simplest applications, such as email, which are non-real-time, and are of very low bandwidth, are pretty easy to offer. But the applications range all the way up to collaborative, interactive video services such as TelePresence. In order for a service provider to offer that to an enterprise or SMB, they need to have the proper network infrastructure to be able to deliver on the SLA [Service Level Agreement] that’s required by the most sophisticated of services. An application such as TelePresence really ‘raises the bar’ in terms of what the network infrastructure needs to deliver.”
“As for TelePresence, Cisco is obviously investing pretty heavily in it,” MacDonald said. “What TelePresence gives you is as close to an in-person experience from a meeting perspective as you can get without actually being there in person. You have a life-size screen, the audio quality is excellent and you have collaborative tools which enable you to share material, presentations, and so on. I’ve been fortunate enough to be in on a couple of TelePresence sessions, and I have to say as a user who has experienced it, I can absolutely see the value that both service providers and businesses will get out of it. I’m scheduled to participate in a TelePresence session tomorrow morning with some folks in Europe, and that saves me a whole week of traveling. I don’t have to get on a plane and spend three days traveling for a two-hour meeting. It’s an important meeting that can’t be done over an audio conference. But of course the experience has to be really positive for this to work and for SPs to get the value out of it. You’ve got the feeling that you’re there, and you get as close to those other participants as you need to in order to have an effective meeting.”
“Even so, there’s got to be a lot of complexity in the network to make TelePresence happen,” he said. “It means that you’ve got to be able to partition the right amount of bandwidth. You’ve got to be able to do it in an environment where you’ve got all of those other services running, with lots of voice traffic and lots of email flying around. You’ve got to make sure that you’ve got the correct Quality of Service [QoS] attributes. You need to be able to handle failure scenarios. If, at the end of the day, the network doesn’t deliver the proper experience, then the end users won’t get the value.”
“My experience with videoconferencing systems was never all that positive,” MacDonald said, “and I really didn’t know what to expect when I first encountered TelePresence. As it turned out, it was a little like seeing the Grand Canyon for the first time for me. It didn’t disappoint and I was really surprised at how life-like it is.”
Virtualized Routers and Blades
“Our recent announcement concerned a service virtualization for the Cisco XR-12000 Series Router,” MacDonald said. “As part of that announcement, there’s a new piece of hardware called the Multi-Service Blade that you can plug into the router’s chassis and it supports multiple virtual-managed services. The first services that we’ll offer on the blade are an integrated firewall and a session border controller.”
MacDonald said the blade has been architected in such a way that it’s extensible to other services.
“So, in the future, you may see us add things such as DDoS [Distributed Denial-of-Service] attack protection, deep packet inspection, lawful intercept capability, and so forth,” he said. “We talk about this ‘virtual capability’; the real value here is that you can take the CPU and memory and then allocate them as required by the services. That’s the real value. What that does for our service provider customers is that it allows them to take the physical resources and to partition them as required by the services that they need to provide for the end customers.”
“The other innovative thing that we’ve done with this blade and the services is that we’ve fully integrated the services into the router,” MacDonald said. “What we mean by ‘fully integrated’ is that once a service goes inside the router, it ‘inherits’ the characteristics of the router. For example, it takes advantage of the QoS mechanism. As you’re probably well aware, we at Cisco have invested heavily in the 12000 Series VPN capability, so immediately the new service can be added to a feature-rich VPN environment. Also, the XR-12000 is designed from the ground up to be a carrier class device. We’ve invested quite a bit in the operating system, IOS XR, to be carrier class in every way. So again, when those services now become integrated into the router, they inherit those characteristics and this increases the SLA quality for the services. This is important for something like TelePresence, which we know will be immensely popular as long as the end user has a positive experience. So TelePresence will sell if the end user doesn’t run into any problems — a high level of integration and high availability are critical.”
“The other key characteristic that the service inherits once it is integrated into the router is that it’s now fully ‘route-aware,’” MacDonald added. “This means that if you have any kind of failure in your network and you must have the data path rerouted around the failure, the service will automatically perform the reroute and it will carry all the characteristics of the service. It basically tags all of the attributes of the service and they’re restored when the service comes back after the failure. Again, you inherit all of those characteristics and investment that happens in the network infrastructure on top of the service, and that provides a more predictable service and a better experience for the user. What it means to the service provider is that now they’ve got a more predictable performance, which is a critical factor in the success of the service, and it also lowers the provider’s overall OPEX (News - Alert). That’s because you’ve moved from a model that has multiple devices to a model where there’s a single device so it’s easier to configure and it’s simpler to manage.”
Appliance-Based vs. Integrated Service Delivery
To get a better handle on the benefits of Cisco’s Service Virtualization ideas, MacDonald discussed the differences between appliance-based and integrated service delivery methodologies.
“Services today can be delivered with various appliances,” MacDonald said. “Generally you have a router and multiple appliances for services. Typically, there’s one appliance per service, and if you need to scale the services, you can rack-and-stack the appliances. What all this means to the service provider is that now he’s got an environment where there are multiple devices, multiple operating systems and multiple configurations. That’s pretty complex. You’ve got to configure the router and configure the appliance, and the appliance typically has static routing capabilities, which means that it doesn’t have that ‘route-aware’ capability that we talked about with regard to the 12000 Series. The SLA that will delivered in this type of configuration is driven by the lowest common denominator in the system. So, if one device has a lower SLA, and that would typically be the appliance, then the SLA for the whole service will be tied to the reduced quality of that particular appliance.”
“On the other hand, if you look at Cisco’s fully-integrated solution,” MacDonald continued, “you now have to deal with only one device. It becomes simpler to configure and manage. More importantly, it inherits the characteristics of the router. Our Cisco routers are designed to be carrier-class from the ground up, so the service takes on those attributes such as route-resiliency. The net is that you’ve got something that delivers a better experience and lowers the total cost of ownership.
How Customers Use These Services
There are various ways that customers can use these services. First, MacDonald discussed how the integrated service border controller in Cisco’s new Multi-Service Blade will enable such things as a better TelePresence service.
“One of the neat things about this service virtualization concept is that we can take the resources and we can allocate them as required,” he said. “As I said previously, you can take the CPU and memory and then allocate them to a service based on the business needs. Imagine that we take a Multi-Service Blade, configuring multiple TelePresence sessions using the integrated session border controller to do the call control and set up the session. One of the neat things that our session border controller has, working in conjunction with the router, is that it has advanced mechanisms whereby it can do a network look-ahead and determine whether the resources that are required for resource-intensive services such as TelePresence are available. And then it takes advantage of all the infrastructure components that already exist in the router. So you can apply the correct quality of service to ensure that the service is delivered effectively, you can put the proper advanced security mechanisms on top of that and, as we’ve said several times now, it ‘inherits’ the high availability capabilities of the router and you achieve the kind of SLA that’s required for this type of advanced service.”
“The Multi Service Blade could then be configured for a different service,” he said. “In another example, let’s take a look at one of our customers, SAVVIS, a service provider in the U.S. Midwest that specializes in global internetworking solutions. SAVVIS has been very focused on achieving as much virtualization as possible. Their desired model is to do as much as they can do in the network, and that gives them great advantages from an operational efficiency standpoint. Currently, they offer their services based on an appliance-based model, but they’re now moving into the service virtualization model where they’ll get everything integrated into the network. We at Cisco are in trials with SAVVIS. They’ve got three of their customers on-tap to do this: GLOBALconnect, Zoom Information and Loyalty Management Services. They’ll all coexist in the same environment. The resources can be partitioned for multiple customers. And again, all of the same things that we’ve been talking about apply: They get to take advantage of all of the feature richness of the router, the QoS capabilities, the route-awareness, the resilience, and you inherit all of those high availability capabilities that we’ve invested in from an infrastructure perspective.”
“Overall, once again they’re able to deliver a better SLA to their end customers and they’ve got a simpler operation environment,” MacDonald said. “In the existing model used by SAVVIS, they have a pretty long time-to-turn-up on these services. That’s because they have one group that manages the router, another group manages the appliances, and yet another group devoted to capacity planning. All of these groups have some input into the system. They’ve got to do configuration changes and all sorts of other activities. In our Cisco model where we’ve got the services virtualized in a single device, however, you still need those separate groups to do their work, but they can now all work with a single device and the coordination among the groups becomes much less complicated. Indeed, SAVVIS’ experience with virtualization is that, for services in general, they’re able to bring down their time-to-market and the total costs of operations on those services.”
Cisco’s virtualization and hardware upgrades will bring greater flexibility, quality and high availability to the world of both service providers and the customers they serve. The future services landscape may appear quite different from the fragmented, appliance-ridden world that now exists.
Richard Grigonis is an internationally-known technology editor and writer. To see more of his articles, please visit his columnist page.