AS IT INFRASTRUCTURE has risen to new prominence in the age of the
Internet, so too has the role of the Chief Technology Officer at
network consulting companies. In his role as CTO of Callisma, George
Orlov helps define strategy for customers who are grappling with issues
of bandwidth, security, directories, and convergence. In an interview
with InfoWorld Editor in Chief Michael Vizard, Orlov discusses these
four networking trends and the impact they will have on the
enterprise.
InfoWorld: What's the biggest trend in the networking segment of the
industry?
Orlov: Everyone is moving to optical. It's a disruptive technology
that allows you to provide pure, raw data bit-rates for 1/100th the
price of your current T1, T3, DS3, or OC3 line. If you invest in this
new technology, you are able to substantially undercut the competition.
If you're an existing customer or existing service provider [and] you
don't enable your network to provide these vastly larger data transport
rates at a vastly lower cost, you will get decimated.
InfoWorld: How will this affect customers, as opposed to service
providers, at the enterprise level?
Orlov: Most enterprises don't have private networks, although it's
getting to the point where if they're in a campus environment it's
cheaper to lay their own fiber than to continue to buy point-to-points
from the existing service providers.
InfoWorld: How well are most people dealing with security and
networking issues?
Orlov: No one's shop is as secure as they want it to be, nor is
security as far up on their radar screen as it needs to be. It's like
earthquake insurance or earthquake preparedness. You never really think
about [being insured or prepared] until the earthquake happens, and
then everyone is scrambling around saying, 'Why didn't we prepare?'
InfoWorld: Do most customers adopt a hide-in-plain-sight security
strategy and then hope for the best?
Orlov: No one overtly says, 'We're going to hide in plain sight.'
But security is much more complex than buying a firewall and plugging
it in. Many people say, 'I've got a firewall, now I'm safe.' Well, 75
percent or so of breaches come from insiders, and a firewall doesn't
help you [in those instances] at all. It's layers of policy, process,
and other things -- long before you buy the hardware or the
intrusion-detection systems or the managed services -- that are
important.
InfoWorld: Is the real threat the enemy within?
Orlov: There are people who get upset for whatever reason and take
advantage of data resources, and they do an immense amount of damage.
Most of that damage is done inside, and much of it is malicious in
nature.
InfoWorld: Why is security so complex?
Orlov: With complex rule sets, you need to understand in detail the
applications and the network layer. It's incredibly complex because
there's not just an IP port. It's about which ports, which applications
use which ports, and which other ports [are] across your different
security zones within your organization. You want to keep those
segmented, because if you hack into one, you get all of them. The tools
for [managing] this are arcane and complex.
InfoWorld: Why not simply outsource this function?
Orlov: I think that for some segment of companies outsourcing your
security and bandwidth management makes sense. There will always [be]
companies whose technology and capability are both sources of their
strength and of their competitive advantage. They need to be nimble
with that technology so that they can get in the contractual
arrangement.
InfoWorld: How important are directories?
Orlov: I think the big challenge for the adoption of Active
Directory on Windows 2000 is that most people who have existing NT 4.0
applications are going to find that Windows 2000 is not an enhancement
of NT 4.0. It is totally different. The security model is different,
and the migration from NT 4.0 to Windows 2000 from the server side is
extremely complex and challenging. I think people will go to Windows
2000, and I think they will go to Active Directory. The time frame when
you will see that uptake will be longer than we had hoped.
InfoWorld: Why is this the case?
Orlov: Microsoft has always designed their directory systems and
their operating systems for enterprises that aren't as big as many [of
the larger] enterprises out there. They still don't sell and target
their solutions to the really big enterprises. Unless your network is
designed incredibly well and you understand that Active Directory sits
on top and is fully integrated with DNS, I don't believe that the
toolsets, ... planning, or design expertise that Microsoft gives you in
their white papers fully prepares you for what you're really going to
find. It's a complete redesign.
InfoWorld: What's your take on Novell?
Orlov: It is great technology [and] works really well, but [it has]
insignificant market share.
InfoWorld: Are any other major trends in play?
Orlov: There's a big drive for convergence. What you've seen over
the past two years is the stalling of old-style telephony systems in
anticipation of things to come. There will be a time in the not too
distant future when you no longer have two sets of wiring plans and two
sets of provision circuits. There's no IT shop that doesn't have two
departments -- voice and data -- but those people don't know how to
talk to each other.
InfoWorld: What's the holdup?
Orlov: The hardest thing is the datacom guys think they're going to
do convergence with the telecom guys. But in order to really do
convergence well, the datacom guys have to fully understand what SS7
switches [for managing telephone calls] do. Meanwhile, the telecom guys
don't want to learn the datacom side because they're a little older and
they've become entrenched. They don't necessarily think that they need
to learn.
InfoWorld: So competition among the telephone companies isn't going
to accelerate this process?
Orlov: Many of the RBOCs [regional Bell operating companies] roll
out DSL only because they see competition from the CLECs [competitive
local exchange carriers]. But by the way, any CLEC that rolls it out
still has to pay the RBOC, so it's not that much competition. Most of
those DSL CLECs are losing money and the RBOCs are still making money
hand over fist. This is because CLECs haven't been able to provide
anything but data services over [DSL]. If they could provide voice and
video and a bunch of other things over the DSL stream, they would start
making some money off of it. The competition on data is going to get
just incredible eventually, but it still costs a ton of money to dig a
hole down the street and lay fiber up to your building.
