| The market for 2.5G and 3G services is
| |
| | applications before deployment may make
|
| taking shape as mobile providers offer
| |
| | success itself a risky proposition. After
|
| customers capabilities that include
| |
| | all, an application that performs
|
| Internet browsing, chat, multimedia
| |
| | perfectly well with several hundred
|
| downloads and online purchasing. These
| |
| | simultaneous users may not survive the
|
| services promise to generate significant
| |
| | onslaught of tens of thousands. So,
|
| incremental revenue for providers, as
| |
| | without appropriate preparation,
|
| customers embrace the next-generation
| |
| | providers can find themselves unable to
|
| mobile experience. They will also be
| |
| | cope with the very success they were
|
| important to other market beneficiaries -
| |
| | hoping for!
|
| including content providers and online
| |
| | The Solution: Network-Aware Development
|
| merchants - who see the affluent mobile
| |
| | One solution to this problem is to make
|
| user as an important target.
| |
| | the real-world conditions of the
|
| To ensure the success of their 2.5G and
| |
| | production network a central design
|
| 3G services, however, mobile providers
| |
| | consideration for all applications from
|
| have to grapple with three primary risk
| |
| | the earliest stages of design. In other
|
| factors. First, they obviously have to
| |
| | words, to ensure service quality in the
|
| ensure that the services they offer are
| |
| | real world AFTER deployment, it is
|
| the ones buyers want. Only by
| |
| | essential to be able to discover any
|
| understanding customers' wants and needs
| |
| | potential performance problems with an
|
| can providers offer a portfolio of
| |
| | application's behavior in the real world
|
| services that will generate the necessary
| |
| | BEFORE deployment.
|
| take-rates to be profitable. Second, they
| |
| | This can only be done, however, if
|
| have to design those services to appeal
| |
| | developers have some practical means of
|
| to users despite small screens, slower
| |
| | modeling those conditions in the lab.
|
| access speeds and limited keypad
| |
| | Ideally, such a modeling environment will
|
| functions. These design skills are
| |
| | be able to replicate conditions on the
|
| essential for optimizing the user
| |
| | production network - including bandwidth
|
| experience and market acceptance.
| |
| | constrains, the number of distribution of
|
| Third, providers must ensure the
| |
| | end-users, etc. If this modeling
|
| reliability of these services. It's one
| |
| | environment is then connected to the
|
| thing to tolerate a temporary outage on a
| |
| | actual server infrastructure that will
|
| free or bundled service. It's quite
| |
| | support the planned applications, the
|
| another thing to lose access to or
| |
| | behavior of those applications in the
|
| experience chronically sub-par
| |
| | real world can be accurately observed and
|
| performance with a service you're paying
| |
| | tested. A variety of "what-if" scenarios
|
| for. That's why it's absolutely critical
| |
| | can also be generated and observed to
|
| for mobile providers to validate the
| |
| | determine their impact on service levels:
|
| performance of their 2.5G and 3G services
| |
| | the loss of some portion of the network,
|
| before they roll them out to customers.
| |
| | an exceptional peak in service
|
| Application failures are something mobile
| |
| | utilization, a spike in other types of
|
| customers are unlikely to be very
| |
| | network traffic, and so forth.
|
| forgiving about. In fact, too many
| |
| | In addition to accurately replicating
|
| failures too early in the game may
| |
| | existing and projected real-world network
|
| permanently turn customers off to these
| |
| | conditions, an effective modeling
|
| money-making services-dooming the market
| |
| | environment should also lend itself to
|
| before it has a chance to take hold.
| |
| | collaboration between application
|
| The Service Level Assurance Challenge
| |
| | development teams and network
|
| Ensuring service levels for 2.5G and 3G
| |
| | infrastructure managers. By sharing a
|
| services presents special challenges.
| |
| | common modeling platform, these two
|
| Multimedia applications, for example, are
| |
| | groups can reach consensus about issues
|
| particularly bandwidth-hungry and are
| |
| | potentially affecting service levels - as
|
| easily prone to session breakdowns due to
| |
| | well as the optimal remedy for such
|
| temporary congestion in the network. In
| |
| | issues - rather than getting bogged down
|
| addition, such applications often use
| |
| | in conflict and finger-pointing. A
|
| multiple session, signaling and data
| |
| | multi-purpose modeling environment also
|
| protocols. These complex dependencies
| |
| | ensures that any investment in modeling
|
| often make it difficult to maintain
| |
| | technology can be fully leveraged across
|
| service levels and understand the issues
| |
| | the entire application lifecycle,
|
| that may compromise the end-user
| |
| | including change management, capacity
|
| experience.
| |
| | planning and CapEx projection.
|
| In fact, providers often find themselves
| |
| | Ensure Risk-Free Wireless Deployment
|
| rolling new services out into production
| |
| | No provider can afford to have their
|
| without fully understanding how those
| |
| | early wireless service efforts tarnished
|
| services will really be experienced by
| |
| | with chronic poor performance. And no one
|
| customers under real-world conditions.
| |
| | wants to surrender their hard-earned
|
| That's because services are typically
| |
| | early adopter customer to the competition
|
| tested in laboratory environments that
| |
| | because of service failures. Mobile
|
| don't adequately reflect the bandwidth
| |
| | providers who want to gain and sustain
|
| limitations, distance-driven latencies,
| |
| | marketshare in value-added wireless
|
| and capacity contention that exist in the
| |
| | services must therefore make every effort
|
| production environment. This is
| |
| | to ensure the reliability and performance
|
| unfortunate, because it exposes providers
| |
| | of their offerings. And to do so, it is
|
| to the considerable risk that
| |
| | critical that they embrace modeling
|
| applications won't perform in production
| |
| | technologies that are as advanced and
|
| as they did in the lab.
| |
| | sophisticated as their next-generation
|
| By the same token, the inability to
| |
| | networks.
|
| adequately assess the performance of
| |
| |
|
