Mobile data traffic is increasing at an enormous rate, which is driving many mobile operators to switch from a 3G network to all-internet protocol (IP) network, such as LTE, to manage the high volume of traffic. The Diameter protocol, which is used on most IP networks, plays a central role in the management of 4G LTE and IMS networks and 3G charging and policy deployments. TelecomEngine spoke with Jason Emery, Director of Product Management at Tekelec, to discuss how its Diameter Signaling Router (DSR) product supports multiple networks, including LTE, through the centralization
Mobile data traffic is increasing at an enormous rate, which is driving many mobile operators to switch from a 3G network to all-internet protocol (IP) network, such as LTE, to manage the high volume of traffic. The Diameter protocol, which is used on most IP networks, plays a central role in the management of 4G LTE and IMS networks and 3G charging and policy deployments. TelecomEngine spoke with Jason Emery, Director of Product Management at Tekelec, to discuss how its Diameter Signaling Router (DSR) product supports multiple networks, including LTE, through the centralization of routing, traffic management and load-balancing tasks.
TelecomEngine: How is the Diameter Signaling Router (DSR) applicable for both LTE and 3G networks?
Jason Emery: The Diameter Signaling Router has five different categories of use cases, and of the five use cases there are three that are particularly applicable to both LTE and 3G networks: charging proxy, policy proxy and core routing.
The charging proxy functionality is a load balancing, topology hiding and session binding capability that operators place in the middle of their offline or online charging solutions. Topology hiding masks the address of the end server from the server that’s connecting to it. For example, if a gateway has a subscriber attached to it, it doesn’t need to know the address of a charging server in the network. It only has to know the address of the DSR. The DSR then picks a charging server to serve that gateway and by doing that operators reduce opex because they don’t have to constantly update all of the gateways in the network with the charging server addresses. It also does load balancing, which is picking a server to deliver the charging information. The last task in this use case is session binding, which makes sure that once a charging server is selected, all the billing records about an individual session go to the same charging server. This is important to eliminate a lot of post processing of billing information that would otherwise have to happen.
Policy control is a Diameter-based function that lives in 3G and LTE networks, and we provide a policy proxy, which has much of the same characteristics as the charging proxy. The difference is that there is actually a 3GPP specification for what’s called a policy Diameter Routing Agent (DRA). This includes things like session binding, loading balancing and topology hiding.
The third use case is core routing. With 3G, LTE and IMS networks having so many Diameter-connected end points, operators will face network congestion, overload, link failure, or other kinds of network issues. Without a core Diameter routing solution, operators would have to rely on all the endpoints to route Diameter messages. But by putting a DSR at the core of the network, they provide a single place to implement network management functions and eliminate the interworking and support issues associated with implementing those functions in edge nodes.
In the case when an operator is partnered with a provider that does not offer LTE, how does DSR support 2G/3G roaming?
JE: One case is where operators simply are deploying devices in their LTE networks by provisioning them in HSSs and don’t want to pay for legacy interfaces, like MAP, on those devices. When subscribers roam off of the LTE network onto another carrier’s 3G network, the only type of roaming interface is MAP. By terminating MAP on the DSR and converting it to Diameter within the DSR gateway function, however, operators don’t need the MAP interface – and its associated costs – on the HSS.
Another example is if operators decide not to establish MAP roaming agreements with carriers that are moving to LTE, instead moving to only Diameter. This would be the case when a 3G subscriber roams onto a neighboring carrier’s 3G network and an operator has simply not established a MAP-based roaming agreement with that carrier because it has decided to move everything to Diameter. In that case the operator can terminate Diameter-based roaming messages and covert them to MAP in the DSR.
Now, in the case of a roaming agreement, how does DSR help with interoperability?
JE: The definition of a roaming agreement includes the notion of interoperability. What’s important about DSR is its ability to mediate between different protocol variants, implementations and different types of technologies.
For instance, the DSR can interoperate between TCP and SCTP and IPv4 and IPv6 connections. We can also mediate where a carrier has one implementation of Diameter and another carrier has a different variant. Some of the requirements we have found from carriers is that there is a set Diameter AVPs that are mandatory, that have to be provided in every Diameter message, but there’s also optional AVPs that can come in messages. The DSR mediation capability allows for the removal of optional AVPs to ensure only the mandatory AVPs remain in the messages.
Growth in number of subscribers usually causes a network operator to expand its network, but this can pose problems. With the increase of subscribers on an LTE network what challenges are presented, and how does the DSR solve them?
JE: The first thing is that the network elements, like the HSS, get bigger. As they get bigger they become a potential single point of failure. Many times carriers will choose to segregate subscribers in HSSs and have multiple instances of HSSs. This way, they don’t have a single spot where a large population of their network could lose service if that network element failed. To do that, operators have a couple of options: they can provision within all the other network elements the location of those subscribers and which HSS has which subscribers; or they can introduce a subscriber location function (SLF) within the network. We provide the SLF on the DSRA subscriber location function allows operators to grow, scale and manipulate the location of the subscribers within HSSs without having to manage provisioning all the information across all the other network elements.
Another factor is when carriers grow through acquisition, they often have multiple HSS vendors in the network. It’s therefore important for the subscriber location function not to be with a point solution on the HSS.
The other challenge associated with scaling the network is the increase in the individual servers that grow to handle capacity. So equipment like charging servers, policy servers and all the devices in the network exponentially grow as the number of subscribers grows. Therefore, carriers often will need load balancers and topology hiding functions to cost-effectively scale.
How does DSR enable policy and charging rules function (PCRF) scalability and solve PCRF binding problems and why is this important to a network?
JE: In the 3GPP specifications for policy control there is a function that’s defined as the policy diameter routing agent, or DRA function. That function is necessary to scale policy networks to provide load balancing and topology hiding capability, but more importantly is what’s called policy session binding and correlation.