It’s not the Americans who are in the box seat to rule the world, it is a cheap and cheerful, somewhat beige technology - IP
Earl “Psst, do you want to buy some TCP/IP” Speranza is a US sales executive who visited Australia about 10 years ago. Even then the subject of IP was old hat, and Earl’s passion for the topic earned him a good ribbing from some quarters. But IP persists - and in fact has risen to a position of great influence. IP is now a cornerstone of the global network infrastructure; it is the common thread that binds every networked device on or off the planet. IP is the technology that is even breaking down the concept of the enterprise network. Today there exist only tributaries.
People are excited about IP. IP is living proof that even the most beige performer can become enticing when associated with enough power.
An analogy that explains the renewed excitement around IP could be some black sludge that came to dominate industry and politics in the century before last. From the 1860s until the turn of the 20th century, oil was the boom-bust industry. From the 1960s until the turn of the 21st century, IT has been the boom-bust industry. Oil’s true role was not realised until WW1 and the rise of the internal combustion engine. That was when we figured out how to best use this resource. The result gave the Allies a decisive advantage on the battlefield, and in the world of commerce delivered a competitive edge to fast-moving industrialists.
Today’s IP-induced excitement is borne on an expectation that history might be about to repeat. Is there an IP-powered internal combustion engine waiting in the wings? And if so, the CIOs who pick its arrival early will deliver their organisations a decisive edge of the magnitude gained by oil-savvy industrials a century ago.
But where to look in a landscape littered with technology trends?
Where else but right at the heart of the Ethernet/IP hegemony. You do not need statistics from Gartner or IDC to tell you what is going on. Fifteen years ago, geeks celebrated when they gave a Coke machine an IP address. Today that’s humbug. There is nothing amazing about a Coke machine with an IP address. There is nothing amazing about anything with an IP address. The infrastructure debate is over. Everything is now connected to everything.
“More visionary individuals see IP as the ultimate protocol enabling all other service offerings and forming the basis of the new, next generation networks,” says a recent Accenture report. If you have any doubt, think again of the analogy and the impact of oil on WW1.
Arguably the Americans’ Ethernet/IP-based BattleNet is a next generation network. And there is little doubt that in Iraq it delivered a competitive edge at least equal to that the Allies gained through oil engines in WW1.
So, what is it we are destined to do with this great new resource?
Shifting to IP?
Solomon said there is nothing new under the sun. Follow this logic and the obvious applications for our ubiquitous Ethernet/IP network pops out. We just shift all our existing communications tools onto it.
Voice over IP (VoIP) has been just around the corner for a decade and may just be starting to make an impact. Gartner estimates that Cisco’s VoIP business has doubled in the last nine months - but off a very low base. “The previous three years were just hard yakka,” says Geoff Johnson, vice president and research director with Gartner. In 1997 Paul Keating, claimed by some as a direct descendent of Solomon, launched an international dialling VoIP service with great fanfare. The assembled IT media was breathless with excitement.
No one else was.
The service never really materialised. VoIP was theoretically possible . . . but. For many organisations VoIP still lacks a compelling business case. Sure, you can save some money by using your data network to shift voice around. And many do; VoIP backbone links connect branch office PABX systems. But to install VoIP to the desktop you have to spend up big installing new infrastructure.
The issue is quality of service (QoS) protocols. Your older switches have to go, and there needs to be end-to-end vendor uniformity as each vendor’s implementation of QoS is just a little bit different. And then you run the risk of swamping your network with voice traffic at just the moment some critical data traffic needs right of way.
Does VoIP offer compelling competitive advantage? Is there a compelling argument for VoIP that goes beyond the net present value (NPV) of future incremental call cost savings? Do those cost savings really stack up against the efficiency of a POTS (plain old telephone service) operating over a network that has long been either fully depreciated or is considered a 100 per cent sunk cost? Could you convince your board that your company will lose momentum in the market without VoIP? Or that your organisation will zoom ahead with VoIP?
For most, the answer appears to be no. Yes, VoIP is being taken up, but most often only as an option to be considered against major PABX upgrades - and even then, POTS often wins. Even so, major branch offices are being linked by IP - with a VoIP plug-in off the back of the PABX. It is little different from what has long been the practice with frame relay. Today, VoIP appears not to be a breakthrough technology offering sufficient benefit to disrupt a preordained upgrade cycle.
Another obvious service is video streaming or rich media applications. On-demand video has an advantage over VoIP in that it is not an existing service already offered on a perfectly capable network subsidised by years of capital investment and billions of dollars in fully depreciated infrastructure.
But rich media applications face two major barriers. First is the “so what” factor. Yes, I can get video to the desktop, integrating my computer, TV, cable connection and DVD player. But I don’t want to watch TV on my computer and I certainly don’t want the staff watching video on theirs. The second barrier is bandwidth. Video is a bandwidth hog. If you run video and VoIP over an existing data network, QoS protocols almost guarantee that data communications will suffer badly.
Potential killer application videoconferencing is still almost exclusively an ISDN dial-up technology. The global IP network is years away from being able to competently handle full-screen, high-definition, ad hoc, point-to-point video communications. It appears that VoIP and video will slowly seep onto the IP network, but without the help of some new, urgent, competitive application, the take-up will be slow. Evolutionary, not revolutionary.
“It’s an unnatural act to put voice and video over a packet network,” Gartner’s Johnson says.
At the turn of the 20th century, automobile pioneer Mercedes was likewise downbeat on the future of the car. It saw a global shortage in chauffeurs as limiting the market potential of its product. Is a global shortage of applications or bandwidth going to likewise limit the future of video over IP? Ultimately Mercedes was wrong. But right now the barriers against full consolidation of communications onto a single IP platform are real enough. The legacy systems work, IP networks are not ready for the task, and a compelling reason to change has yet to surface.
Was Solomon Wrong?
Unless Solomon got it wrong, the final ascendancy of the IP hegemony is still a fair way down the track. Although, perhaps it is not something new we are waiting for, but rather something that we are yet to notice - waiting to emerge and push the technology along at a faster pace.
IP pundits are talking up a trio of logical data technology services - virtual private networking (VPN), distributed storage area networks and distributed processing.
Virtual private networks are not a new idea. What is new is the increasing elegance of their implementation and function. It is hard to judge what is driving the popularity of VPN technology. One could argue that it is the increasing reach of the IP hegemony to include ever more devices of ever greater variety that is driving the VPN value proposition. Is this the same as arguing we did not evolve arms until there was something worth reaching for?
An alternative argument is that security concern is driving the adoption of VPN. The increasing reach of the IP hegemony means there is more to protect against, more reason to define and protect your own “space”. This argument has extra clout when you consider the nature of many enterprise networks. Being IP networks, just like the rest of the Internet, are they themselves little more than VPNs, tributaries defined by some security processes?
Whatever the case, VPNs today reach to increasingly varied and useful resources against increasingly complex business and security processes. Yes, VPN is just a tool. In any operations environment, maintaining control of a process requires finite boundaries to be placed around it. VPNs create logical boundaries around an organisation’s network. This allows the implementation of rules within that defined area; security rules, workflow rules, access rules.
As a tool, VPNs make a lot of sense. The killer application of the future may well be driven by an increasing dependence on VPN technology to link us to the resources we require. The board that was so unmoved by pleadings regarding VoIP may change its tune when faced with the possibility of using VPNs to manage multiple directorships from a single IP device.
It is not just humans that can leverage VPN technology. Storage area networks (SANs) were originally conceived of as dedicated server farms: rooms full of disk drives.
Distributed SANs mean that your server farm can be just as distributed as your user base. In theory you can store your data anywhere, hunt the Internet by night and quickly fill up people’s disk drives before they notice. And what you can do for storage you can do for processing as well. Grid computing is a re-emerging trend - resource sharing is back in vogue. As useful as these logical services are, just like VoIP, none offer more than incremental benefit.
Logical services are primarily about resource sharing. Better utilisation of disk space or processing power. With storage and processors now cheaper than ever, the benefits of improved efficiency can be insufficient to justify the administrative overhead.
This year the IP network became quite literally universal. At least two of the four spacecraft now en route to Mars have IP addresses. In October 2003 the closest will be about 50 million kilometres away. The connection is obviously wireless.
Wireless is one segment of the networking industry where genuine technological breakthroughs are taking place. The business case for wireless is compelling. Any portable device can now send and receive data.
A key driver is that data can be much more efficient than voice in terms of resource requirements and accuracy of transmission. Taxi operators now run thousands of cabs off a single channel and can be certain that each driver receives and understands the exact address each time. This compares to the past where a voice channel could not handle more than 60 to 80 bookings an hour, and for each booking there was a chance the driver misheard the address.
Courier company Mail Call recently invested in excess of $3000 per vehicle, for more than 200 vehicles, ripping out voice and putting in a wireless IP device running over GPRS. The payback is almost instant - based solely on the licence fees for the radio spectrum no longer required.
Another emerging technology is wireless broadband: portable, private, secure, high-speed broadband connections from anywhere within five kilometres of a base station. This is potentially a disruptive technology, undoing the need for cable or ADSL infrastructure.
Wireless technologies such as GPRS present a certain irony. IP pundits are extolling the virtues of shifting circuit-based technologies, such as POTS, across to packet-based networks. However, GPRS does the opposite. It takes a packet-based technology - IP - and runs it across what is from a logical perspective, a circuit-based network (GSM).
Packet-based networks have always been based on the assumption that latency was tolerable. But latency is not tolerable for voice or for real-time video (video streaming is okay because of buffering). This in turn creates the need for the development of workarounds (read QoS standards). But most of an IP network’s efficiencies are inherent to the trade-offs it makes against quality of service.
By implementing QoS you reintroduce elements of the inefficiency inherent in circuit-based networks. You do this by reserving bandwidth for certain applications, or at least giving some traffic a decided priority.
The solution is to throw a tonne of bandwidth at the problem, so that regardless of the demands placed on the network, it will deliver. But this in turn presents another clash. If the rationale sitting behind the integration of all communications onto a single IP network was increased resource efficiency, you are here undone. In order to meet base line performance levels you need to invest in a great deal of bandwidth that will rarely be used.
To this IP pundits have a final retort: IP allows you to measure exactly who is sending what across the network at any given time. At peak times, when resources are expected to be scarce, people who use bandwidth-hungry applications can be penalised.
It sounds very much like a utility business model - which is what it is. The IP hegemony means the homogenisation of what was always destined to be a utility service. It will not be that long before Homer Simpson represents the skill level required to administer most enterprise networks.
Where’s the Party?
Sad, isn’t it, that after 30 years of fantastically exciting fast-paced innovation we are left with something as boring as a seamless, homogeneous utility service.
And there is the nagging doubt that IP has won not so much because it is the ultimate protocol as because it is the default protocol. Everyone employs it because everyone employs it. There is no real reason or driving force. There is no compelling argument beyond vanilla operational efficiency.
But, while we all stand around kicking the ground with our hands in our pockets, there is still a sense that something is going to happen.
A Broken Nexus
There is an alternative view of the future, one where networking technologies fragment again into many different directions. Special purpose networks are being talked about; networks purpose built for just video, or just voice, or just data.
Ben Wrigley, IT manager of Hotel InterContinental Sydney has done just this. There are four networks, each quite separate, in his hotel. First, there is the administration network, then the phone network and the movie network. Finally, there is a broadband IP network connecting each guest room to the outside world. In theory, all four networks could be consolidated into the one, but Wrigley has gone to some effort to ensure some of them do not even come into physical contact with each other.
Underpinning Wrigley’s decisions is a firm belief that many applications are as incompatible as oil and water and are incapable of existing on the same network. This of course flies in the face of the latest buzz to emerge from the US - content delivery networks (CDNs). These are structured to allow all content varieties to coexist seamlessly on the one network. There are two main arguments being used to propel the adoption of this technology.
Gartner’s Johnson uses the “it’s inevitable” argument. End users are quite simply going to demand access to voice and video to the desktop. All those CIOs out there who are busy building regimes that exclude everything but simple text e-mails will be forced to buckle. Vendors like Cisco use the “distributed applications” argument. Gosh, applications will be that much more powerful with all this stuff - so let’s get going with it.
It’s the Business, Stupid
When CIO Connected contacted Steve Adair, director IT for Mail Call, the interview was short. Adair was busy preparing a tender.
The tender was for the provision of services beyond the “norm” for a courier company. It was for a company that wanted to outsource an internal business process. Mail Call was leading with a bid supported by its state-of-the-art technological capabilities. In fact, Adair’s deployment of technology over the past couple of years has become central to Mail Call’s competitive position. He has achieved this through the implementation of a distributed application .
The resolution of much of the IT infrastructure debate releases many CIOs from a technology focus, allowing a much greater emphasis on a business focus. Simultaneously, the utility and reach of the IP network means that applications impossible to consider yesterday will be easy to implement tomorrow. Using VPN technology running over a CDN you can tie together any conceivable combination of resources into a single operation.
From here on it is up to us to drive the technology. The universality of the IP stack means network technology has become fully commoditised. The creative spirit is what will drive the future of networking, not so much in regard to thinking up better networking technologies, so much as figuring out what we should be doing with it.
This trend is not happening before its time. Executive decision makers are demanding more business savvy from their technology teams, especially the technology leadership.
We will know for certain that the IP future has arrived when we stop talking about it. After all, IP is nothing if not vanilla. It is a building block, the sort of technology that we should all be able to take for granted. The action will be increasingly around the question of what to do with it. We have this cheap, cheerful, ubiquitous network ? is there something we can do with it to really drive our organisation forward? Can we be the first to think up some new way of leveraging this resource to great competitive advantage?
If the analogy fits, and networking today is where oil was exactly 100 years ago, then we are on the cusp of a very exciting time. If only we can figure out what to do with it.
Case Study | Mail Call
The Call of the Wireless
The digital age hasn’t been kind to couriers, but one company is using the global IP network to get its own back
Couriers have had it tough.
First the advertising industry jumped ship - sending ads in electronically rather than physically. The six o’clock rush from advertising agencies to the newspapers vanished. Then the business document market dried up. E-mail was to blame. Then the legal document market started to wither. Now only bulky items, or items bearing an all-important signature, require couriers.
You would think the IT director working in the courier business would be singularly unpopular.
Not so Steve Adair. His company Mail Call has turned his skills at utilising the global IP network infrastructure into one of its core competitive advantages - and while other courier companies wither and die, Mail Call is going from strength to strength. To Adair it is simple - everything is now connected to everything. Customers can book courier services using an IP device - their desktop or a PDA - and a driver can receive that booking via an IP device - a PDA.
Mail Call used to have 120 staff handling bookings. Now there are just 47 for that purpose.
But the biggest saving is in bandwidth. Mail Call had multiple voice channels, each licensed and attracting huge annual fees. Now they have none. All communication with the fleet is managed over a public network. This opens expansion opportunities. All Mail Call requires to enter a new market is a couple of drivers. There is no need to license bandwidth, no need to set up transmission facilities, no need to build a call centre.
Adair has built a true distributed application, one that encompasses the entire operational processes that underpin the business.
The one downside is a loss of some of the team spirit. Drivers used to hear all the bookings and the banter that went with it over their two-way radios. That’s gone. “Everyone would hear someone make a mistake, and know not to do the same thing,” says Adair. “Now everyone makes the same mistake - there is a loss of collective intelligence.”
Case Study | Hotel InterContinental
Welcome to the Hotel Well-Connected
One hotel in Singapore has learnt that broadband connections bring in the customers
The very lowest time on any business trip is that time spent downloading e-mails onto the laptop in the hotel room. It normally happens at 3.00am, you’re jetlagged, there is driving rain outside and there is no one left alive anywhere on the planet to offer any technical support.
Four- and five-star executive hotels are starting to figure this out - and sense the opportunity for some competitive advantage. Gartner’s people spend a lot of time on the road, and in Singapore they stay exclusively at one hotel purely because each room has a broadband Internet connection.
Ben Wrigley, IT manager of Sydney’s Hotel InterContinental, became an early mover when presented with a golden opportunity - a complete refit of every one of his hotel’s 540 rooms.
Now a fibre/cat 5 network runs to every room offering ‘wire speed’ connections to the Internet via a MAN socket in the hotel’s basement. The entire network is physically separate from the hotel’s administration network. And Wrigley resisted the urge to consolidate movie and voice traffic, leaving existing dedicated circuits in place. “Although I am sure some guests are using the IP network to bypass our telephone service,” says Wrigley.
Wrigley was sorely tempted to shift the movies across, but in the end decided against it. “It would have been a decision based on a desire to show off some technical prowess, not based on any sound business case,” says Wrigley.
The potential bandwidth demands that could be generated if each room moved to wide-screen high-definition screens were staggering. But, as an insurance policy, Wrigley has placed an extra network port behind each TV screen.
The biggest issue for the network was security.
Lots of IT executives stay at the hotel. If the guest network was connected to the hotel’s administration network, and a technology-savvy guest was awake all night jetlagged, Wrigley is unsure where they would end up. But even with the network totally separate, there is the security of the other guests to worry about. “We had Al Gore staying here a few weeks ago,” says Wrigley. “You couldn’t have a system whereby other guests could snoop on his hard drive.”
Thus the implementation includes 540 virtual networks - one for each room. It is all managed by a mysterious black box in the basement. “When you get to hundreds of VLANS you get to the limit of what the infrastructure can handle,” says Wrigley. That Wrigley will not give CIO Connected any details on the box speaks to the competitive advantage he sees inherent in his 540-VLAN network.
The next step for the hotel is wireless. All public areas will be converted into hot spots. Guests or visitors will be able to order a coffee and half an hour of wireless broadband - and charge it back to their room or a credit card.
And finally, there is a help desk that Wrigley swears will be available 24 hours.