Guest

Cisco IP Transfer Point

A Study in Mobile Messaging

Table Of Contents

A STUDY IN MOBILE MESSAGING:
The Evolution of Messaging in Mobile Networks, and How to Efficiently and Effectively Manage the Growing Messaging Traffic

Executive Summary

Historical Overview: A Quick Look at Existing Messaging Architecture

Messaging Evolution in 3G Systems: SMS, MMS, and SIP

The Need for a More Efficient Architecture

Market Trends

Conclusion

White Paper—spring 2004

A STUDY IN MOBILE MESSAGING:
The Evolution of Messaging in Mobile Networks, and How to Efficiently and Effectively Manage the Growing Messaging Traffic


Executive Summary

Mobile messaging technology is evolving rapidly to provide multiple services and applications to today's subscribers. With the continued delay of third-generation (3G) deployments and migration to fully packet-switched networks, coupled with the demand for fast-to-market data applications over existing wireless networks, mobile operators look more than ever at their short messaging systems to fill the gap.

Existing Global System for Mobile Communications (GSM) and Code Division Multiple Access (CDMA) mobile networks use Short Message Service (SMS) as a multipurpose data service that enables rapid deployment of data applications without the need for 3G bandwidth capabilities. In fact, SMS traffic is growing exponentially, straining existing mobile operators' signaling networks and infrastructure. The original intention of the SMS protocol was to provide the mobile subscriber with information about the mobile network condition. SMS has recently become a popular protocol for text messaging over the airways.

Traditional messaging deployments are based on a centralized model where all messages sent and received by the mobile user are typically directed to messaging centers within the operators' networks. This model is becoming less effective due to the increased traffic and usage levels of messaging in the mobile networks today. The centralized model causes excessive messaging loads on the SMS centers (SMSCs) as well as irregular bursts of messaging traffic that is typically sent over the traditional signaling links, such as Signaling System 7 (SS7), originally designed to handle voice signaling and not the bearer data traffic. A more efficient architecture is necessary to alleviate this problem. Today, there are two different yet complimentary approaches to enhancing the mobile messaging environments:

The first approach is to provide a reliable and efficient offload of signaling traffic to the more cost-efficient and easier-to-manage IP network. The next-generation architecture proposed by both the 3rd Generation Partnership Project (3GPP) and 3GPP2 call for the use of Signaling Transport (SIGTRAN) technology in the future deployments of such mobile networks. SIGTRAN uses translation protocol layers to take traditional SS7 traffic from a circuit-based network to a packet-based network, essentially the IP network.

The second approach is to provide a simple message-handling function, usually referred to as first delivery attempt (FDA), at the edge of the mobile network. This is essentially a messaging-center function designed to intercept and attempt the delivery of the message.

These two approaches complement each other by providing improved messaging-traffic handling and efficient offload and delivery of messages.

Historical Overview: A Quick Look at Existing Messaging Architecture

In today's GSM and CDMA networks, messaging is based on basic SMS or the enhanced Multimedia Messaging Service (MMS). Of these, SMS is the most prevalent in mobile networks because of its ease of use, earlier adoption, and its availability on older mobile stations (handsets) such as second- and 2.5-generation handsets. Both of these technologies are implemented in a centralized architecture where the mobile device or application sends a message to the messaging center. The messaging center then locates the mobile device and delivers the message to that device. This store-and-forward approach helps ensure the mobile device will receive the message even if it was offline at the time the message was sent. Figure 1 illustrates an example of the existing messaging environment.

Figure 1

SMS Delivery—Traditional Model

SMS control and data messages are sent via the traditional circuit-based voice signaling, or SS7, network. MMS, however, uses the packet network introduced in 2.5G and 3G environments by delivering the content over the bearer network (General Packet Radio Service [GPRS] or Code Division Multiple Access CDMA2000). Note that in MMS, some of the control messages still traverse the signaling network.

Current architecture enables mobile operators to scale their messaging capacity by increasing the capacity of the messaging center. Capacity is also increased by using more or larger messaging center platforms as well as by adding signaling bandwidth in the network (circuit and packet).

The majority of messaging traffic can be classified into three categories:

1. Mobile originated-mobile terminated

2. Mobile originated-application terminated

3. Application originated-mobile terminated

The messaging services and applications deployed over any mobile network today can be classified into one or more of these categories. A television voting campaign is an example of mobile originated-application terminated traffic. A sound clip or animation sent from one mobile network to the other is a mobile originated-mobile terminated type of message.

All traffic in the above categories must traverse the messaging center. Peak application use, as might occur during an interactive TV program, for example, can cause a traffic spike that results in serious consequences for wireless networks as messaging centers reach their capacity limitations. Other events, such as sporting events or holidays, can have similar effects on wireless networks.

Messaging Evolution in 3G Systems: SMS, MMS, and SIP

3G promises enhanced services and much higher bandwidth for end users. However 3G deployments continue to be delayed. To date, there has been no significant deployment of many promised multimedia services—such as videoconferencing, interactive multimedia sessions, and streaming content—in the mobile world. With this continued delay, operators with 2.5G systems look to existing messaging infrastructure for solutions.

SMS and MMS are being used as bearers for some of the applications that don't require the 3G level of bandwidth. Voting, instant messaging, gaming, and other applications are being used today over SMS. Short messages with animated multimedia are often transferred over MMS.

The evolution toward an all IP-based infrastructure remains the ultimate goal of both operators and vendors. The obvious benefits of such migration include reduced capital and operational costs, as well as enhanced services for the end user.

However, the 3G evolution is exactly that—an evolution. This means that reaching the all-IP architecture may take longer than initially anticipated. Most operators will migrate slowly toward this architecture.

From a messaging point of view, client-to-client environments will be more prevalent in 3G systems. Session Initiation Protocol (SIP) provides this environment. With clients registered at a centralized server, other clients can request a location resolution to establish an SIP session with each other. SIP can carry voice, video, and messaging traffic, thus providing a flexible transport layer over IP that results in better end-user services and experience.

Simply because of the efficient architecture that makes SIP a better messaging bearer than SMS (which is SS7-based), or even MMS, SIP-based messaging will eventually become more mainstream than SMS or MMS.

The Need for a More Efficient Architecture

The preceding observations outline a single commonality among the different messaging technologies: that is, the centralized model used in most messaging implementations. A more efficient and enhanced architecture may use a hybrid model that combines a centralized model with a distributed function.

Each time a user sends an SMS message, for example, that message must route to the SMSC. Today this causes a bottleneck scenario during high-volume use of SMS, as occurs when audience interaction, voting, and gaming applications are used, or during seasonal events and holidays, which can cause tremendous spikes in messaging traffic. In certain cases, some operators experience loss of service for extended periods of time simply because there is not sufficient capacity in the messaging centers to handle the traffic. Figure 2 illustrates today's voting environments.

Figure 2

Voting Traffic—Traditional Model

A more efficient architecture can be realized in which a centralized messaging center is used in conjunction with distributed messaging intelligence at the edge of the network. A basic message-handling function must be installed at the remote MSC sites, intercepting the messaging traffic and making intelligent routing decisions to send the message. For example, if two people are located in a city served by a single MSC and one sends a message to the other, the message will be intercepted at the local MSC site and delivered without having to go through a centralized server. This capability can be used to enhance numerous types of other services as well.

Following is a list of different types of SMS traffic with descriptions of how the distributed messaging function helps enhance the delivery of such traffic:

Voting, gaming, and audience interaction—This is usually referred to as mobile originated-application terminated. In this scenario, the voting or gaming traffic is sent directly from the distributed messaging function to the voting server or the gaming server, thus bypassing the message center. This results in more efficient delivery of the voting traffic without overloading the SMSC.

Push service, location-based service—This is typically application originated-mobile terminated. Currently such traffic would have to be sent to the SMSC before delivery to the mobile device. This can be enhanced by sending the traffic directly to the remote messaging function for delivery.

Mobile-to-mobile traffic, or mobile originated-mobile terminated traffic—In this scenario, the messaging traffic would be isolated in the remote site (a metropolitan area, a state, etc.) without having to traverse the core WAN to reach the messaging center.

The preceding examples list different types of traffic that may be efficiently handled and routed at the edge of the network without having to traverse the core network or reach the messaging center.

A distributed, intelligent message-handling function would, however, have some limitations. These may include limited disk space, limited processing power, and limited logic and processing capability. Therefore, such capability is typically used for a simple delivery attempt called First Delivery Attempt (FDA) to keep the requirements to a minimum. This means that if the destination of the message is not available to receive the message, the FDA server would then deliver the message to the centralized server for deferred delivery. However, more than 60 percent of mobile-terminated messaging is successfully delivered on the first attempt. The benefits of this approach far exceed any limitations. Figure 3 illustrates the FDA function distributed to the MSC locations.

Figure 3

Efficient Messaging Architecture

A distributed messaging function performing FDA clearly provides an efficient enhancement for handling messaging traffic. Additionally, this capability may provide core-capacity savings by offloading SMS traffic from the SS7 network. This could be accomplished by using IP-based messaging technology and protocols such as Short Message Peer-to-Peer Protocol (SMPP) between the FDA function and the messaging application servers. Undelivered traffic may also be sent to the messaging center for deferred delivery via such protocols or via SIGTRAN-based connectivity. SIGTRAN provides efficient and reliable SS7 transfer over IP. 3GPP and 3GPP2 both call for the use of SIGTRAN technologies in 3G+ networks.

Market Trends

Operators today are actively searching for network enhancements and cost-savings measures to increase profitability. As messaging becomes an important form of communication in mobile networks, mobile operators need more efficient ways to manage their messaging traffic. Most operators are seeing tremendous benefits by deploying an FDA function at the edge of the network.

Many operators are migrating to a SIGTRAN-based messaging offload as well as a full SIGTRAN-based signaling network. This trend will continue as the benefits of SIGTRAN technologies accelerates their continued deployment in mobile operator environments worldwide.

Conclusion

Existing mobile networks usually include a centralized messaging architecture. In such an architecture, all messages are routed to a messaging center for delivery to their destination. This places a strain on the costly messaging center and on the core network transporting the messaging traffic. As more applications are deployed using SMS and MMS as bearers, and as users become more accustomed to using such applications, operators need a way to offload the traffic from the core network and messaging server.

Distributed messaging intelligence achieves more efficient message handling by implementing an FDA function, thus containing a large portion of the messaging traffic to a specific metropolitan area. Such distributed intelligence provides two main benefits: offloading messaging traffic and using IP-based delivery of messaging and other signaling traffic.

IP based messaging and signaling transport architectures provide mobile operators a more flexible and efficient model resulting in lower capital and operating expenses, as well as increased capacity and performance.