eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Softswitches today are like the PCs of the 90s: their introduction not only improved the way people live and work, but changed the entire infrastructure so that its more efficient and easier to use.
And if the softswitch is the PC, which in many cases it literally is, then the old Class 4 and Class 5 carrier switches of yesterday and today are the mainframes — old, traditional, bulky machines that did a few things really well, but not a whole heck of a lot.
“A lot has changed since the mid-80s in hardware technology,” notes Jack Kozik, former developer of Class 4 and Class 5 switch applications and present director of architecture for Lucent Technologies e-services group, a unit that builds softswitch technology. “It used to be the software had to be robust because the hardware would easily break down. Now, the software doesnt have to be perfect, because even the general-purpose hardware is so reliable.”
As carriers begin to deploy softswitches throughout their network — mostly as additions to Class 5 switches, rarely as replacements — the developers who created the traditional telephony applications that most people use today look with fondness, wonderment and even trepidation at the newer revolution.
A softswitch controls telephone calls from either the traditional circuit-switched networks or the new packet-switched telephone networks — and sometimes both — using standard hardware and innovative software, which can be easily updated. Softswitches also interact with application servers to add features such as call waiting or call forwarding.
The biggest difference between softswitches and carrier switches is the openness between separate platforms, Kozik says. Traditional switches were extremely proprietary, from the hardware to the operating system, the software, the programming language and the debugger. As a developer, Kozik was required to learn the difficult programming language and develop in it for months upon months, often going more than a year, until an application emerged.
Since U.S. customers expect nothing short of perfection when they pick up the phone, the applications were stringently tested, and then finally deployed very slowly over a number of years. In some cases, it could take up to four years before an application was widely deployed.
“This presented a major challenge when you wanted to develop two applications for the same switch,” says Dave Barclay, a marketing director in Lucents e-services business unit and former developer of Class 4 and Class 5 switch applications. “The challenge was interaction between the two, making sure they interacted the way they were supposed to. From a service creation standpoint, this was big.”
And the same people would have to develop those applications. A carrier couldnt ask one programmer to develop call waiting while another developed caller ID, because the programs might not work on the same machine.
It required months to master the languages in which programmers were developing, whereas todays softswitch applications are done in C++ or Java, which most computer science students learn in the first couple of years of college.
Net Heads Catching On
But that doesnt mean that the hotshots coming out of school today have anything on the old guys. If anything, they have much to learn. “We call those guys the Net Heads, and were the Bell Heads,” says Steve Donovan, a software architect of Dynamicsoft. “The Net Heads are those kids coming out of school that learned to develop for the Web, while us Bell Heads, of course, are the older ones that developed for the large carrier switches.”
Donovan started out developing telephony applications in the 80s for the telecom group of Digital Equipment, now part of Compaq Computer. He then moved on to MCI in the early 90s to develop some of the first wireless applications. Finally, Donovans interest in Session Initiation Protocol (SIP) led him to Dynamicsoft a few years ago.
SIP is the underlying protocol for next-generation applications; it looks a lot like HTML, and allows programmers to develop applications in simple languages, such as Java. Dynamicsoft, one of SIPs early adopters, creates application servers and applications that will run on the softswitch architecture.
“People coming out of college today dont have any idea what it was like back then,” Donovan says. Mostly, they dont understand the underlying hardware the way that older developers do, so they cant appreciate why an application has to be developed a certain way. But theyre catching on, he says.
The older programmers arent having an easy time either. “Anybody coming from the circuit-switched world had a difficult time embracing the IP world,” Donovan says. “Theres just a different mind-set. Before, if you got one byte wrong in a message, it was rejected outright, even if it wasnt important to the application you were generating. In the SIP world, theres a different approach. Even if the message is corrupt, you can still send the parts that arent.”
But what impresses Donovan about softswitch technology is very similar to what impressed Kozik and Barclay: the open platform. In the 80s, to even be considered a telephony developer meant that you were employed by a major equipment manufacturer that had a direct relationship with a carrier, such as Bell Atlantic or Pacific Bell.
This severely limited the potential for great applications, because not only did you have to know the language and the specifications, you needed one of the monolithic switches on which to test your software — and you couldnt pick one up at the local electronics store. Todays softswitch applications are Web-based and only as complex as IP itself, which means that some eager beaver toiling in his or her garage could develop a voice application in a few days, using just a PC and an Internet connection.
In the early 80s, John Weald was a developer at Bell Labs, where he worked on application processors — devices designed to enable adding features to circuit switches easier. Today, Weald is vice president of software engineering of softswitch maker Sylantro Systems in Campbell, Calif.
Weald says that what he likes about the new softswitch applications is the ability to open them up to consumers to manage for themselves. Previously, the only way a subscriber could manage something like voice-mail was to call up and key in some changes to the greeting and ring count.
The new applications coming out today enable the subscriber to go on the Web and tell the system where he or she is going to be at certain times of the day, and instruct the system on how to handle each call. So, if someones boss calls at 9 p.m., he or she can tell the system to have the call directed to his or her home phone, and follow up with a call to a wireless phone if the boss cant get through.
Weald says that this is a 180-degree turn from the days of the closed-off proprietary development environment in voice services. “Now, you have the browser on someones PC as the interface to these applications,” he says.
But what seems to bring the most pleasure to the older developers are the innovative applications on the horizon that could never have been thought of before.
Lucents Kozik and Barclay like the location-based services, such as using a wireless phone to find the movie theater nearest you — not by punching in a ZIP code, but because the system knows where you are at all times. Parents could give their kids wireless phones and, as a nice proviso, they get to track their childrens movements using the Internet.
Weald likes to rely on the Net Heads for the new ideas. While they may not appreciate what it was like “back in the day,” they do have a knack for dreaming up the applications of the future.