|
Networks & Infrastructure
Tekelec gets to the core of the NGN
SSR core router application alleviates softswitch deployment pain
by Sean Buckley
Migrating to a Next Generation Network (NGN) certainly has its benefits of enabling the service provider to expand their respective VoIP and multimedia desires, yet like any network migration, the NGN is not without its drawbacks.
One obvious drawback is that the NGN does not have a core signaling and session control layer. The lack of these layers limits the service provider's ability to scale their softswitch deployments.
Tekelec, however, believes the answer lies in adopting a centralized network core.
Traditionally, when a service provider deployed a NGN with softwitches, they had to provision the routes at every softswitch. All of this information would then have to be distributed to every softswitch in the network. This creates a logical mesh of routing information where every softswitch knows who is supporting who, which becomes more painful as the operator starts to grow their network.
“Softswitches and MMS serve both a signaling component and the router component for data,” said Sridhar Karuturi, Product Manager for Tekelec. “The way the networks have been now connected is that every softswitch is literally connected to every other softswitch at the logical layer. Every softswitch is provisioned with information on which subscribers are being served by what softswitch.”
Instead of building direct SIP signaling connections to every softswitch, Tekelec’s EAGLE SIP Signaling Router (SSR) creates a centralized core session framework.
This new framework allows for the service provider to offload session management duties off of the endpoints. In addition, the session layer can implement various IMS benefits to the NGN including access independence, multimedia support, Home Service Control (HSC) and subscription-based service orchestration.
New core competency
A common opportunity yet problem service providers are facing as they migrate to an NGN network and adopt an IMS architecture is the ongoing growth of VoIP.
Subscribers have tuned into VoIP for not only lower-cost calling, but also web-based functions (on-screen voice mail) and its portability (users can take the VoIP provider’s MTA device to any location and make calls and be charged only for local calling rates).
VoIP’s popularity has required service providers to continually deploy softswitch nodes to keep up with growth. Because a service provider’s softswitch is based on a TDM switched overlay where routes and SIP trunks are predefined, every softswitch has to use a TDM algorithm to translate the destination number into pre-defined routes and SIP trunks.
Adding to the pain is the fact that every route and trunk has to communicate subscriber information to every other softswitch in the network. Perhaps not a major problem for a small carrier with only a few softswitches deployed, but for large carriers that are deploying multiple softswitches this construction will hamper their ability to scale gracefully.
“There’s a logical mesh of routing information in the sense that every softswitch knows who is supporting whom,” Karuturi said. “That becomes extremely painful as you want to grow the network. Let’s assume there are 10 softswitches in the network and you want to add the eleventh one. When you add the eleventh one, you not only have to do what is needed to add the eleventh one, but you have to go back to your existing 10 and say what are the new routes the eleventh one is going to support.”
Serving as the central routing point for the softswitch, the SSR will perform the function of a session manager to extend centralized to either the softswitch (wireline) or the Mobile Switching Center (MSC) servers in a wireless network.
figure 1. SSR creates a SIP-based reference architecture
Unlike the traditional routing model, the softswitch will now route all calls by default not to other softswitches but back to the SSR. The SSR can then make Layer-5 decisions on advanced algorithms and forwards the requests to the right SIP destination.
“What we’re saying is centralize all the routing so that way when you add a new softswitch nobody else in the network has to know about it,” Karuturi said. “When a call arrives at a softswitch and if the softswitch does not serve this particular subscriber all it does is route that information to the centralized server which is the SSR. That way when you grow the network, only the SSR has to know.”
Given the tight economic times, any dollars a service provider can squeeze out of their existing network equipment the better. The SSR not only will enable the service provider to get more capacity out of their existing softswitches, but also avoid having to pay additional charges for SS7 interfaces for number portability when they purchase softswitches. Softswitch vendors typically don’t include this functionality in their base packages. (see Cable-Tec Expo 2008: Tekelec relieves IP peer pressure and Local Number Portability hits its stride).
“We’re seeing the softswitch vendors charge for an SS7 INAP interface for number portability,” Karuturi said. “Buying that interface on 10 different softswitches now, you basically have the SSR at one point and that’s all you need.”
figure 2. Softswitches, serving as PoPs, form a fully meshed architecture.
In addition to shaving off additional costs, the SSR can act as a central monitoring point so service providers can monitor the calls all from one location.
Connecting the endpoint dots
Hand in hand with the growth of VoIP, service providers are seeing residential customer’s appetite for broadband services (video, voice and gaming) show no signs of slowing down.
To access those services, a growing set of SIP-enabled gaming devices and Set Top Boxes for IPTV are coming to the forefront.
With many of the devices accessing these services being SIP-capable, a service provider is finding themselves deploying edge and core proxy servers so multimedia can interface with media servers, for example.
Service providers have traditionally had two options to get their edge and core SIP proxies: leverage free ware or public proxies or using softswitches. The drawbacks of these approaches are that free ware lacks stability, while using a softswitch limits the ability to route non-voice traffic. Just as it can serve as a control point for softswitches, the SSR can also function as a core or edge SIP proxy.
In this application, the SSR registrar function accepts registration requests for SIP endpoints, provides authentication services, and hosts the subscriber’s address of record (AOR). It can also accept SIP requests for regulated SIP uniform resource locator or user profile.
Thus far, Tekelec said carrier interest in the SSR is growing especially in Asia-Pacific, Europe and in Brazil. Currently, a tier one provider in Europe that has two dozen MSCs in their network has completed a technical trial of the SSR, while another operator is using it for number portability over SIP.
|
