Why LTE Needs Diameter Signaling Management

By TMCnet Special Guest
Ben Volkow
  |  April 29, 2013

As LTE (News - Alert) networks are rolling out worldwide, there are many discussions going on about the critical need to manage signaling, in particular Diameter signaling. There are also discussions, and it’s even fair to say some controversy, over the different methods to manage and control Diameter signaling in the most effective way.

Before we get into the nuts and bolts of Diameter management in LTE, to understand Diameter protocol better let's look at legacy signaling protocols in the early days of the addition of IP to networks.

In 1980, the signaling protocol SS7 was introduced by the International Telecommunications Union to control telephone call sessions through point-to-point connectivity. Only after scalability and management issues became apparent, the need for centralized management was clear and a new network entity called the signal transfer point was introduced to manage, connect and route SS7 traffic.

SIP, the communication protocol to support voice calls over the Internet, has similar origins. Originally SIP was intended to connect network entities point-to-point, which lasted a while until management, interoperability and routing requirements evolved and the session border controller was introduced in 2003 to handle and solve SIP management issues.

Beginning in the 1970s, the data plane was also initially designed on point-to-point connections. Few people imagined that there would be so many connections, and much data and signaling traffic to render this architecture obsolete. But, as we know, data and signaling traffic increased, which expedited the need for switches, routers and load balancers to support signaling traffic management and scaling.

When Diameter was first introduced by the the Internet Engineering Task Force, it included the concepts of Diameter agents that can proxy, route and balance Diameter traffic to provide scalability and management requirements. However, when the 3GPP promoted Diameter as the foundation for signaling in IMS and EPC architectures, it left Diameter Agents out, perhaps thinking that the introduction of a distributed architecture would avoid the need for Diameter signaling management.

But history repeats itself, and the same issues that affected legacy signaling protocols arose around Diameter in IMS and EPC. As a result, the 3GPP, the mobile broadband standards organization, went back to the drawing board and proposed the use of a Diameter Agent to support Diameter signaling traffic. The group named it Diameter Routing Agent and agreed that it's needed around the PCRF (which stands for policy charging and rules function).

Today we are witnessing the same Diameter issues that arose around the PCRF, now appearing in other network locations. Therefore, we believe it won't be long before the 3GPP will issue another announcement around Diameter management enlarging the scope of the DRA’s network deployment.

Six Reasons Supporting the Need for Diameter Signaling Management

Here are six reasons why it is imperative to deploy Diameter Signaling management solutions in an LTE, or any IP-based network. This list is not exhaustive but reflects the primary reasons.

The growth of Diameter signaling

Operators are moving from a voice-centric to a data-centric environment. Unlike voice, which did not generate much signaling during a session, data creates a lot of signaling. Furthermore, the amount of signaling associated with data sessions differs from application to application, and depends on configuration and usage patterns. In some cases signaling is up to 50 times higher in a data session in comparison to a traditional voice session. We know from customers that some operators are experiencing a monthly growth of Diameter signaling of 25 percent. With this drastic and fast growth, operators find themselves plagued with the required management and scalability requirements needed to prepare their networks before the inevitable Diameter tsunami hits and brings down their networks. Gradually, they are arriving to the conclusion that they must deploy Diameter traffic management solutions in their networks.

The fragmentation of networks

Networks are much more fragmented today than in the past, meaning that they are built with many more boxes. This occurred because the 3GPP introduced new concepts for IMS and EPC that evolved from the definition of physical devices to conceptual functions to fulfill more finely-tuned requirements to handle policy management, session control, charging, media and other functionalities such as I-CSCF, P-CSCF, S-CSCF, vPCRF, hPCRF, MGCF and others. As a result, today we have almost twice the number of defined network boxes than we had a decade ago.

Of course, this fragmentation dictates the need for additional signaling interfaces and connectivity, which results in much more signaling traffic and complexity. Another trend that influences the fragmentation of the network is the growing number of data centers intended to reduce costs and bring the network closer to the subscriber.

However, the increase in data centers requires more signaling connectivity across and between them. It’s clear why this growing fragmentation requires much more signaling management to control the increased number of network functionalities, sites and interfaces for complete reliability and assurance in network performance.  

The growth of usage patterns

We have many more usage patterns today than operators would encounter in the past. When we primarily used our phones to make voice calls and simple data usage with three to four well-defined services like WAP, MMS and minimal Internet surfing, keeping in mind that these were temporary connections to the data plane - the user had to open a new PDP session every time he or she wanted to use data and once finished, the user would close the connection.

This is nothing like today’s smartphones and all other net devices usage behavior through which users are constantly connected. This constant use means we are sending signaling messages all the time. In addition, each application presents different signaling behavior and usage patterns, both of which are serious management challenges to operators. 

The need to support new use cases

The over-the-top trend is making life difficult for operators. The growing market competition and customers’ freedom to easily move between operators affect revenues. To make up for this loss operators have turned to service differentiation, offering compelling new packages and services to subscribers, such as family plans, shared data plans, policy, QoS and so on. All these new services require signaling solutions for support of the service itself and to support the additional signaling created during its operation.

There are a few more trends related to the growth of use cases that increase signaling volume. Another one worth mentioning is the growth in pre-paid subscribers both in developing and developed countries. Pre-paid billing creates more signaling compared to the traditional postpaid billing.

Every new use case requires close management of the Diameter signaling in the network, and moreover, requires context-aware signaling solutions that can access a wide array of databases to enable those services and applications.

Diameter signaling characteristics and the move to all IP networks

Unlike legacy signaling protocols, which were predominately circuit-switch based, Diameter protocol is always packet based and uses TCP or SCTP as a transport protocol to enhance reliability. However, TCP creates twice as much network traffic due to the need to ACK all messages (meaning every message must send a receipt message). Sending a receipt automatically doubles the number of signaling messages.

As clearly seen, the move to an all-IP network significantly increases the amount of signaling. Although the move from circuit switch to packet switch will bring many other advantages, it generates tons more data traffic, and data is one of the major forces behind growth in signaling. Operators need to confront the blitz of signaling from a multitude of fronts never seen before and must be managed before damage is caused to their networks.

New Specs and Trends

LTE and the EPC core offer fresh architectural concepts and are made possible thanks to Diameter signaling protocol. Diameter is completely different from the existing legacy signaling protocols. Due to its relative young age and unfamiliarity in the market, Diameter challenges operators to architect and secure their networks properly.

Since the beginning of LTE rollouts, we have already witnessed a few Diameter signaling related outages around the world, and operators can't afford any others.

But the challenges do not end with Diameter knowledge in a typical LTE network. We have completely unprecedented emerging trends and technologies such as virtualization, software-defined networking, VoLTE and machine-to-machine technology. Each one of them introduces its own architectural concepts, and each generation of technology brings different Diameter behavior patterns and requirements. Together they present daunting tasks in designing, scaling and safeguarding networks.

VoLTE, for example, represents a move from traditional circuit switch-based voice to packet switch data-based voice, and subsequently due to the nature of IP, generates massive signaling. This results in quality of service issues that must be handled by

Diameter signaling management.  

M2M represents the introduction of signaling-centric networks on a scale never seen before, with billions of devices generating signaling. It's not only the amount of signaling that requires a new way to approach signaling management, but also the different signaling behavior of machines when compared to traditional networks with human subscribers. Just imagine about a million ATMs sending updates to the network all at once or a city-wide system of reporting on waste removal or fleet management.

These new trends require finely-tuned Diameter management, overflow control, signaling routing and load balancing to be successfully launched and maintain network reliability.

Ben Volkow, formerly CEO and co-founder of Traffix, is vice president of product development at F5 Networks (News - Alert) (www.f5.com).

Edited by Stefania Viscusi