December 1998
The Three "Lucky Sevens" -- How We'll Hit The
IP Networking Jackpot
BY TONY RYBCZYNSKI
It is difficult not to be awestruck by the Web. Nothing is beyond its scope - for, in
my mind, the Web includes the Internet, intranets, extranets, and everything to which the
Web connects. Also, it is difficult not to marvel at how quickly the Web has expanded its
domain. Who heard of Webtime a few short years ago?
But, as impressive as the Web may be in terms of breadth, it is perhaps even more
impressive for its depth. The Web is, in fact, positioned to become the communications
infrastructure of the future.
The Web already shares many of the attributes of the current telephone network. In
addition, it has its own unique attributes, and it stands to acquire others, that will
allow it to supersede the existing network. Indeed, technologies are already emerging that
will enhance the Web, and extend its benefits to network operators, enterprise users, and
consumers.
In broad outline, the Web possesses certain attributes (I list seven below), which, in
combination with certain emerging technologies (again, I list seven), will allow the Web
to provide certain benefits (again, seven). Ultimately, all these sevens add up. Just as a
string of sevens may click into place in a slot machine, releasing a torrent of riches,
the Web's string of sevens will realize a vision of next-generation networks. Even better,
with the Web, the jackpot isn't a matter of chance. It is, rather, a matter of design, and
the execution of that design.
THE SEVEN ATTRIBUTES
Here they are, the key attributes of the Web that are positioning it as the communications
infrastructure of the future:
- Standards. With the Web, standardization isn't about just IP, but all
the protocols built around it, a real acronym soup, including TCP, FTP, HTTP, RTP, H.323,
and L2TP - to name a few.
- Omnipresence. The Web is ubiquitous. It is manifest in the LAN, the
campus network, the enterprise, and over the Internet. Yes, there are some desktops that
don't yet support the IP stack, but these represent a shrinking minority.
- Location Independence. The Web shrinks the world, making e-mail
connectivity a breeze and information access virtually location independent. The Web is,
in a word, global.
- Ease Of Use. It's easy to use Web-based graphical user interfaces - so
easy, in fact, that users are sometimes unaware when they are networking.
- Addressing. The Web defines an addressing space that everyone knows in
the form of e-mail and WWW addresses (of the form [email protected] and www.nortel.com, respectively). IP addresses of the form
47.4.55.128 are critical; however, they are, to a large extent, under the covers from a
user perspective.
- Application Support. The Web, as the application platform of choice, is
the basis of all new apps; in fact, there are a million developers out there working on
IP. Yes, there are legacy mission-critical applications that are based on Novell's IPX and
IBM's SNA protocol suite, and these will be with us for years if they get by the
millennium. But there is no debate about new apps.
- Signaling. The Web defines a growing set of signaling protocols to
provide transparent connectivity to get at information and ultimately to connect people.
If you think about it, these same seven attributes make the telephone network the
communications network it is today. It's standards-based, it's certainly ubiquitous, it's
global, it's easy to use, it's got a global numbering plan, it's got lots of applications
riding on it, and it's controlled by a fairly sophisticated signaling system (for example,
SS7). So, what is it that the telephone network lacks? Not much: It's got great
performance (including low latency), it's highly reliable, it's got a tested business
model allowing carriers to bill for service and make money. Fundamentally, when you get
right down to it, there's only one thing the telephone network lacks: bandwidth for bursty
data traffic.
Enter the IP world, which is optimized for data and can support voice and multimedia.
So much for accommodating bursty data traffic. But what does the IP world lack? The
telephone network's fabled reliability. And, not only does the IP world need to become
more reliable, it has to get more class to support mission-critical applications,
including voice. Also, it needs a few more standards (for example, standards to provide
interworking with legacy networks, and to match the feature richness of telephone
networks).
THE SEVEN TECHNOLOGIES
Here's my list of technologies that will enhance the Web, in its broadest sense, for
network operators, enterprise users, and consumers:
- IP Quality Of Service (QoS). It's all about meeting the needs of
applications - the current best-effort model of the Internet has to go! No two apps are
created equal - some are mission-critical, others are very delay-sensitive, and yet others
are bandwidth guzzlers. Two major IETF initiatives to watch: Integrated Services and
Differentiated Services (or IntServ and DiffServ, for short). The former will be used in
enterprise nets; the latter is much more scalable and is destined for the Internet.
Interworking between the two is a must.
- IPSec. This is just one of several security-related standards (Public
Key Encryption is another). These standards are critical if the Internet is going to meet
its potential for e-commerce and virtual private networking.
- Multi-Protocol Label Switching (MPLS). MPLS is being developed by the
IETF and introduces a lot more order into IP networking. IP traffic is mapped at the edge
of the network into Layer 2 label switched connections, enhancing traffic management and
allowing VPN operation to be driven by service level agreements (SLAs).
- H.323. This family of real-time multimedia communications and
conferencing standards are critically important in delivering on the vision of voice/data
convergence and collaborative apps.
- Consumer DSL And Cable Modems. ADSL has been highly touted, but it
faces some daunting deployment obstacles. It's not about local loop limitations; it's
about having to send a technician to each residence to rewire the house. This is what's
behind splitterless consumer DSL - just plug the PC into any telephone jack, and you're in
business. Meanwhile, cable modems have their own deployment issues, but are seeing
increased penetration.
- Next-Generation Wireless. This is all about meeting the needs of road
warriors for high-speed data access. Next-generation wireless standards are still in the
works, but if mobile telephony is any indicator, there is a huge demand for wireless
internet access.
- Dense Wave Division Multiplexing (DWDM). DWDM is part of every SONET
fiber optics system providing multiple wavelengths of light on a single fiber. For
example, with a 32-wavelength DWDM system, a single fiber can support 32 10 Gbit/s OC192
payload or 320 Gbit/s. That's lots of encyclopedias per second! What's happening now is
that ATM and IP switches want to get at this bandwidth directly without the SONET layer,
which is primarily there for the multiplexing of lower speed channels, protection
switching, and maintenance signaling. All this translates in lower bandwidth costs.
That's my top seven. What's yours? Feel free to e-mail me your list. Maybe I'll publish
it in a subsequent column.
THE SEVEN BENEFITS
And now the third tumbler clicks into place. Here's my list of seven benefits:
- Webtone Networks. Continued transformation of the Web into what John
Roth, Nortel's CEO, calls Webtone networks: better reliability, class of service support
to differentiate applications and traffic, new billing models.
- Voice On Whatever. Convergence is a given. Transporting voice over
frame relay, ATM, and (increasingly) IP is a given looking forward. The real value comes
from new IP telephony features that enable new business applications accessible from IP
phones, PC phones, Internet-enabled call centers, and wireless sets.
- Roll Your Own VPNs And VPN Services. This is all about leveraging the
connectivity of the Internet: process reengineering and new inter-partner business apps.
It's also about outsourcing enterprise networks with improved price/performance and better
reliability, and about freeing up resources to focus on leveraging networks rather than
running them.
- Application-Optimized Networking. IP QoS and enhanced security are
coming to an IP network near you. But on their own, these can result in havoc rather than
improved networking. That's where policy management and directory-enabled networking come
in. These will provide needed coupling between end user profiles and networking
capabilities: who is allowed to talk to who; who can request what network resources; who
should be charged for network resources. These come together to deliver
application-optimized networking.
- Multi-Layer Routing Switches. With the importance of IP routing, the
nature of the Web networking engines is changing. Routers are being transformed from
software-driven multiprotocol engines to IP-optimized switching platforms (with packet
forwarding in hardware) to multi-layer routing switches.
Why multi-layer? Layer 3 is IP. Layer 2 is frame relay, ATM, or MPLS, and it is there to
enhance the traffic management of IP networks and to strengthen the delivery of IP and VPN
service guarantees. Layer 4 is moving closer to the application and allows the network to
better meet application needs.
So, expect to see lots of product innovation for the campus, and for enterprise wide area
and public networks. These will be the platforms for application-optimized networking.
- Consumer Always On Broadband Access. PCs on LANs are always attached to
the network and this enables a range of applications from message notification to
multicast news and stock quotes. This model traditionally is broken in the residential
environment, in which case narrowband switched modem and ISDN access is used. This is all
changing to always on access over broadband cable modem.
- Death Of Distance. The Web signals the death of distance, making the
desktop area network almost seamless with the global network.
THE JACKPOT
The cumulative effect of all these "sevens" is to realize a vision of
next-generation networks that eliminate the boundaries between LAN/WAN in latency and
bandwidth, between voice and data, between wireline and wireless, and between service
providers.
Tony Rybczynski is director of strategic technologies and marketing for Nortel's
Bay Network Division. This business unit offers a full range of enterprise workgroup,
campus, and wide area unified networks, through direct and indirect channels. For more
information, visit the company's web site at www.nortelnetworks.com.
E-mail questions or comments to the author at [email protected]. |