No carrier and vendor can escape the facts:
Your product development has changed, is changing or will change. And if
you're talking about change, speed is the name of the game. The PC and
LAN industries learned that long ago.
To deal with these changes, a new business
of process is emerging. Solutions are being developed to help telecom companies
improve their efficiency. A funny thing: Process-based products to help
companies change their product-based processes.
One such solution is the Quantum Leap Methodology
(QLM), developed by systems designer and OEM product developer Kalman Saffran,
president of Kalman Saffran Associates, Newton, Mass.
Upstart sat down recently with Saffran
to discuss QLM and also sought the views of Michael Doss, VP of marketing
and business strategy for Siemens Information and Communication Networks
in Boca Raton, Fla. Doss has never seen QLM but agreed to offer his thoughts
on the concept of a product to improve product development.
Upstart: So what is QLM?
Saffran; "It's a fairIy fundamental, straight-forward
system, not complicated at all. It addresses all phases of the product
development lifecycle, from specification through testing and release.
QLM is a system that consists of several components: One is the methodology
itself, another is a set of processes that implement the methodology. Below
the processes, we have methods and procedures. We also have techniques,
tools, templates and frameworks."
Upstart: But "methodology" just means a
specific way of tackling a problem, right?
Saffran: "That's the essence of it. QLM
consists of a framework of processes, and the processes address the lifecycle
of product development. So there are processes for specifications, for
architectures, for designs, for verifications, etc. When you put all those
processes together in a framework that kind of controls your execution,
you call that a 'methodology.'"
Upstart: So in very basic terms, the idea
is not having to reinvent the wheel every time you develop a new product,
but use reusable building blocks instead?
Saffran: "Yes. It's based on the fundamental
concepts of accuracy, precision and consistency - as well as on the '80-15-5'
rule. In other words, the top 5% of activities contributes 95% of the results.
QLM is a methodology that focuses on that top 5% in particular, then secondarily
on that 15% and doesn't address the bottom percentage at all."
Upstart: Why did you develop QLM?
Saffran: "Companies were taking too long
and spending too much money to develop products. In a lot of cases, after
all was said and done, the products were non-performing or not scalable
or they had too many bugs. In some cases, the company tended to trade off
quality for time. Of course, the winning combination is a high-quality
product developed fast."
Upstart: Just how long is today's time
to market?
Saffran: "The average time to market ranges
in various industries, including telecom, from about 18 to 24 months -
although the computer industry is a little faster these days, probably
12 to 18 months. Most companies want to reduce that by a substantial amount
- in a lot of cases, they want to cut it in half."
Upstart: Do you agree with that estimate
for telecom, Michael?
Doss: "Yes, and we're all feeling the pressure.
That pressure has been there for some time in the enterprise LAN arena,
as traffic has grown amazingly fast in the net-works there. The equipment
becomes obsolete very quickly, and therefore the user is constantly purchasing
new equipment.
"Now it's starting to be a real pattern
in WANs, obviously driven by the Internet [traffic]. Everybody will use
the words 'doubling in less than a year,' and some will be aggressive enough
to say 'doubling every three months.' When you get that kind of growth,
there's enormous pressure on the switches and routers in your backbone
network. What you install to handle your traffic this year just won't handle
your traffic next year, and you've got to install a whole new set."
Upstart: All right then - you speed up
your time to market, cut your costs and become a tougher competitor. Why
can't companies do that themselves?
Saffran: "In the past, efficiency of productivity
wasn't an important issue, Now that there are many new competitors in the
marketplace, it's not just a question of, can you get your product or service
to market, but also how fast can you do it?
"A lot of companies don't have this in
place because they lack internal expertise in methodology. In most companies,
there has not been a formal function to address product development methodology
per se."
Upstart: How do you feel about that, Michael?
Doss: "This phenomenon of faster time to
market is just now happening on the WAN side. So it's really new for all
of us. Without commenting on their capability, if Mr. Saffran and his people
can do what they say they can do, I think they've got a winner.
"Now, I think we will all use companies
like his as tools. We also will develop internal techniques - we'll do
a lot of things. And five years from now, Mr. Saffran probably will have
to find another industry segment that's coming under time-to-market pressure.
[By that time], I suspect the big telecom vendors will have responded to
this, just as I believe the LAN and PC vendors already have responded internally.
I don't think we're nearly as far along as the LAN and PC manufacturers
are, in terms of time-to-market tools and programs."
Upstart: Who's more likely to use your
QLM - carriers or the equipment manufacturers?
Saffran: "Both, really, but probably the
manufacturer more than the carrier, simply because the manufacturer develops
many more products than the carrier does. QLM is applicable to all of a
manufacturer's product development - hardware, software, datacom, telecom,
cable TV. Quantum Leap Methodology-Information Technology, or QLM-IT, is
oriented toward the development of MIS applications in an end-user environment."
"So, let's say it takes a manufacturer
two years to develop a switch. With this methodology, it's possible to
reduce that time by up to 50%, to approach one year."
Upstart: Is it really possible to chop
50% off development time?
Doss: "Yes, and it's a necessity. If you
go back to the thought that Internet traffic is doubling every 4 to 6 months,
that tells you that within two of these doubling cases, your backbone switch
or router is out of gas. The backbone you build today has to be replaced
in less than l2 months, maybe closer to 9 months.
"If you're an ISP, your vendor better have
a new product available 9 to 12 months down the road that at least doubles
or triples throughput and capacity - or you will switch to someone else
who does have that."
Upstart: Besides ISPs, are vendors feeling
the same pressure from CLECs and other carriers?
Doss: "Yes, Most of the typical 'communications
operators' now are actually running four to six net- works - a voice network,
an lSP kind of network and three or four data networks for their business
customers, frame relay, circuit emulation, IP, SNA, what have you. Operating
separate networks is costing them an enormous amount of money. They would
like to converge those.
"This convergence is the most important
thing happening in the telecom business day. We don't think it's going
to get down to one network, but we believe this half-dozen networks will
get down to three: the narrow- band voice network, an Internet net- work
that handles only IP traffic and then this third 'converged' network, which
has to offer high quality of service for all these different data types.
Initially it will be ATM-based, and eventually it will be based on IP technology
- as soon as we add a few steroids to IP to make it strong enough.
"Obviously, the Internet network is the
one that's now driving this time-to-market, but we think that as voice
and video move over to this converged network, it will start growing that
fast, too. This phenomenon of time-to-market is going to hit the converged
network in the next couple of years. So once we all learn to turn out products
in 9 to 12 months, instead of l8 to 24, we're going to have to learn to
do it also in these more complex, higher quality-of-service networks."
Upstart: Will carriers awarding equipment
contracts look more favorably on QLM-based vendors?
Saffran: "Yes, because aside from the reduction
in time-to-market and associated costs, which could be passed on to the
buyer, there's also the predictability aspect. With QLM, you can fairly
accurately predict your schedules and your interim milestones, so the carrier
has a high level of confidence of getting what it's buying in the requisite
time frame."
Doss: "Another point here, just a couple
years ago, what you had to do as a vendor and what the service provider
offered was a fairly narrow set of things. Now our customers, the service
providers, want to provide all these networks, so the breadth of our product
range has grown tremendously.
"One way to fill in your product set -
when you need more breadth than you've got the engineering depth to do
- is OEM. But another way is to use people and tools like Mr. Saffran's
company to leverage your own engineers more. So I think that (with operators
broadening all their service offerings, and equipment vendors [getting]
into a huge range of stuff, you're going to see more of this [outsourcing].
No one company can do it all."
Upstart: What does QLM consist of anyway?
Saffran: "What we call 'Phase Zero' is
a review of a limited operation within a manufacturer, to see where the
areas for greatest improvement lie. For example, we would produce a report
that identifies key areas where there are any potential returns. We would
then rank those returns in a financial way, based on ROI and payback period.
Then the next phase would be applying the methodology to one or more of
those areas."
Upstart: Does QLM face any competition
in the market today?
Saffran: "We haven't done complete market
research to know exactly what the competitive products are, so let me give
you a 'back-of-the-envelope' analysis. Essentially, the products in the
IT domain are for particular aspects of the lifecycle - a product that
helps you do specification, another that helps you do testing, etc. Most
of these are tools, and a tool is a very narrow, limited thing. It's not
a methodology. On the engineering side, there are also many, many tools,
but they don't address the methodology end, either. We've essentially developed
a complete lifecycle system. On the selling side, whether a company is
developing software-based systems or both hardware- and software- based
systems, to our knowledge, there are no competitors."
|