By Amichai Lesser
The market for 3G services is taking shape as mobile providers offer customers capabilities that include Internet browsing, chat, multimedia downloads and online purchasing. These services promise to generate significant incremental revenue for providers, as customers embrace the next-generation mobile experience. They will also be important to other market beneficiaries - including content providers and online merchants - who see the affluent mobile user as an important target.
To ensure the success of their 3G services, however, mobile providers have to grapple with three primary risk factors. First, they obviously have to ensure that the services they offer are the ones buyers want. Only by understanding customers' wants and needs can providers offer a portfolio of services that will generate the necessary take-rates to be profitable. Second, they have to design those services to appeal to users despite small screens, slower access speeds and limited keypad functions. These design skills are essential for optimizing the user experience and market acceptance.
Third, providers must ensure the reliability of these services. It's one thing to tolerate a temporary outage on a free or bundled service. It's quite another thing to lose access to or experience chronically sub-par performance with a service you're paying for. That's why it's absolutely critical for mobile providers to validate the performance of their 2.5G and 3G services before they roll them out to customers. Application failures are something mobile customers are unlikely to be very forgiving about. In fact, too many failures too early in the game may permanently turn customers off to these money-making services - dooming the market before it has a chance to take hold.
SLA challenge
Ensuring service levels for 2.5G and 3G services presents special challenges. Multimedia applications, for example, are particularly bandwidth-hungry and are easily prone to session breakdowns due to temporary congestion in the network. In addition, such applications often use multiple session, signaling and data protocols. These complex dependencies often make it difficult to maintain service levels and understand the issues that may compromise the end-user experience.
In fact, providers often find themselves rolling new services out into production without fully understanding how those services will really be experienced by customers under real-world conditions. That's because services are typically tested in laboratory environments that don't adequately reflect the bandwidth limitations, distance-driven latencies, and capacity contention that exist in the production environment. This is unfortunate, because it exposes providers to the considerable risk that applications won't perform in production as they did in the lab.
By the same token, the inability to adequately assess the performance of applications before deployment may make success itself a risky proposition. After all, an application that performs perfectly well with several hundred simultaneous users may not survive the onslaught of tens of thousands. So, without appropriate preparation, providers can find themselves unable to cope with the very success they were hoping for!
Network-aware development
One solution to this problem is to make the real-world conditions of the production network a central design consideration for all applications from the earliest stages of design. In other words, to ensure service quality in the real world after deployment, it is essential to be able to discover any potential performance problems with an application's behavior in the real world before deployment.