Nobody Said It Would Be Easy

The number of wireless stock-trading services has grown from zero to at least 10 over the past year. For investors, the appeal is the ability to react immediately to the latest tip or trend. For carriers, wireless e-trading is one more service to differentiate themselves and boost revenues as price wars heat up.

But making e-trading and other enhanced services such as caller ID and voice mail as reliable as basic voice service can be challenging. INs are complex to begin with, and enhanced services only add to that complexity. Subscriber profiles are one example.

"The profiles are getting so complicated with all the different services they could have," said Kim Parker, Tekelec product manager, IN-diagnostics division. "Not only the number but the interworking (are concerns): If they have this service, how does it affect that service?"

Assessing an enhanced service's quality of service (QoS) isn't easy. There's no shortage of performance statistics for the SS7 network and individual elements, but raw data doesn't necessarily give an accurate feel of what the subscriber is experiencing.

"We can make some assumptions about what they're experiencing, but there's a certain level of subjectiveness to that experience that a signaling-system (monitor) can't replicate," said Chris Loberg, Tektronix business-development manager. "It's dangerous just to assume that sitting in a central office or switch location, an operator can develop a clear understanding of the customer's experience."

SYMPTOMS, BUT OF WHAT?When subscribers report problems with their enhanced services, the symptoms might point in a different direction from the actual cause.

"You may have no problem with your voice-mail application, but you may have a lot of noise or an MSC problem," said Grant Wakelin, ADC software-systems president. "That's why we look across the network at all the elements."

An end-to-end view of the network provides more than just a way to identify services and subscribers affected. It's also a convenient way to narrow down the potential causes.

"Start at a top-down view to get a perspective on the overall network behavior and the quality parameters that affect that," said Kevin Cavanaugh, ADC software-systems division director of product research. "Some people will dive to the detail too quickly and not be able to make the correlations of all the different aspects of the element behavior and the service performance."

For more in-depth troubleshooting, a protocol analyzer is handy.

"Use a protocol analyzer with capabilities for multiple interfaces with multiple databases, the switch and the subscriber handset on different links," said Martin Polak, Tektronix product-marketing manager for access networks. "The data should be captured on multiple calls. A good protocol tester has features such as call trace and intelligent filter functions to help track down the problems amidst a huge amount of data."

TOO MUCH OF A GOOD THING?Network design can help limit the effects of an outage. One example is spreading multiple enhanced services across multiple signaling control points (SCPs). If too many services share the same hardware platform, managing them can be a headache because a change to one can affect the others.

"You could have one SCP and have 15 services, and that's certainly a nice, cost-effective way of doing it," said Stuart Rosenfield, Alcatel senior director of marketing for WIN. "But we recommend segregating some of the services on SCPs. What you don't want is your prepaid subscribers taking down your caller-ID subscribers. Determine which features are critical and which are not."

The distributed approach apparently is catching on with carriers. "We preached this a couple of years ago, and everybody said, 'You just want to sell more stuff,'" Rosenfield said. "Then they had an accident or their jobs were in jeopardy, and they said, 'Maybe it's not such a bad idea.' Although I don't advocate one service per SCP, I think you have to be smart how you group them."

Grouping could be helpful because the number of enhanced services only will increase. Part of the appeal of WIN and emerging standards such as CAMEL is that they give carriers more flexibility to create their own enhanced services. But increased ability also means more responsibility. Just because a new service can be developed and deployed quickly to stave off competitors doesn't necessarily mean that the know-how and tools to manage that service can keep pace.

"Operators say, 'Boy, I just can't wait to create services,'" Rosenfield said. "But you've got to have training with that. Although it's simpler than it used to be, it's not simple. I think some carriers (believe) that it's like a PC, and it's not. You have a lot of back-office and management issues. It's not trivial."

Vendors also will feel pressure to develop equipment and tools that can accommodate a new breed of services developed and deployed on the fly.

"We usually flow from what's coming out of the standards bodies," said Tekelec's Parker. "Even if it's proprietary, it's proprietary Nortel or Ericsson, so it's usually something that's not specific to an individual operator."

Because each network is different, and each enhanced service would be different, carriers could be left to determine what effect "home-brewed" services will have on their networks. "Not only would they have to create that service, they'd probably have to come up with the equipment to test that service," Parker said.

Then there's the human element. Enhanced services can be bewildering to subscribers, so one helpful troubleshooting tool is the ability to track an individual subscriber down to the keystroke level.

"You have an audit trail of what the subscriber is doing and what responses the application has been sending," said Brad Davis, PulsePoint Communications director of product marketing. "This is a pretty resource-intensive feature, so you can't do it on every subscriber at the same time. We recommend setting it up for representative features of each type of service and for problem customers. It gives them a track to uncover user error and correct it."

Under the hood, enhanced services are even more complex. One unhappy result is what's sometimes called "cockpit error."

"A lot of legacy systems that carriers deal with are Unix-based, and their operational interfaces are cryptic, command-line-driven things," Davis said. "They're very error-prone and non-intuitive. People can think that they're entering the right information when in fact they're not. What happens is you get an end user provisioned incorrectly."

Not good for an e-trader on the go. Overlook QoS, and they could end up with a service that's more Austin Powers than James Bond -- except they won't be laughing.

"More than any time in the past, marketing and engineering need to work hand in hand," Rosenfield said. "You can kill a successful service if they say, 'It never works, so who cares?'"