Quelle: CIO, Kanada
If you're a CIO, one of your biggest concerns is probably'missing the boat': focusing on the wrong technologies ornot investigating the right ones.
Communications is one area where CIOs may feel quite vulnerable inthis regard. There's so much going on in the communications space thatit's nearly impossible to keep up with the various emergingtechnologies.
To help make sure that the boat doesn't sail without you, we talked toseveral experts to clarify some of today's most significantcommunications technology trends.
Converging Lines
When we asked our experts which technologies CIOs should be payingattention to, they all agreed on one: convergence of voice, data andultimately video on one IP-based network. Not that it takes aclairvoyant to see this - the industry hype machine has been ballyhooingIP convergence for years. But now the hype is starting to take on moresubstance.
Ian Angus, of Angus TeleManagement Group, Ajax, Ont., has been an IPconvergence sceptic, but today he says it is "potentially the biggestshift in how we do communications since the move in the 1980s fromdumb terminals to PCs and networks."
The big payoffs will come, he believes, from applications that mergedata and voice functionality. One example is a hotel system thatallows guests to order from room service by pressing buttons to selectfrom menus on their telephone's screen.Another is an inbound sales system that automatically loads acustomer's record onto a PC screen when a call comes in.
"None of these are things that you can't do without IP," he stresses."It's just that IP makes it affordable and relatively easy to do."
Mark Quigley, director of research at Yankee Group, also sees IP voicebreaking through soon. "We've been talking about it for years," hepoints out. "But now we're finally getting to the point where IPnetworks are able to provide the service levels needed - the vauntedfive nines of reliability [99.999 per cent up time]. And now we canstart to see the cost savings and potential increases in internalefficiency and productivity."
Many companies are experimenting with IP telephony in branch offices,notes Quigley, adding that it's a low-risk strategy that more shouldadopt.
At the very least, IP convergence should be part of long-term planningnow, says Angus. Most firms should also now be making moves toconvert, upgrade or expand infrastructure to accommodate futureimplementation of converged systems.
"A lot of the things that companies are definitely going to want to dowith IP convergence in 2005, they won't be able to do if they don't makechanges in 2002 and 2003," Angus warns. "If you haven't thought aboutit at all or you've been continually brushing it off, then you've made amistake and you'll be playing catch-up next year."
In fact, only about 40 per cent of companies have started to make themove to IP, and fewer than one per cent have completed the migration,Angus estimates. Gartner Group data also shows a slow migration.IP-capable systems accounted for 10 per cent of new phone system salesin 2001, up from three per cent the year before. They will account for20 per cent this year.
But there is nothing wrong with that, says Mark Fabbi, vice presidentand research director with Gartner Canada. "We think the adoption rateshould be slow. We don't feel there are compelling application reasonsfor doing it right now. Total cost of ownership [compared totraditional telephony] is break-even at best. In the vast majority ofcases the network needs to be upgraded. And IP terminals are not cheap- we're still in a pre-standard environment."
He adds, "It's clear everybody will be IP eventually, but there is noreason to drop what you're doing today or to shorten the life of anexisting PBX. The roll-out will take a relatively long time."
Roberta Fox, president of Fox Group Consulting, Markham, Ont, is alsosceptical of any across-the-board rush to migrate to IP voice. "Inmost cases, there are no compelling financial reasons for doing it,"she argues. Current discounts for telco voice circuits make itunlikely that IP voice would generate significant savings. There areexceptions, though, she notes - for example, aggregating long-distancetraffic to a centralized call centre over an IP backbone network.
The fact that many vendor proponents come from a data background andhave to fight for credibility in the voice arena is another hold-back.Now that some telcos in Canada are getting behind IP with serviceofferings, companies may have more confidence to move ahead.
Next Up: Video
The immediate impetus for IP convergence is the promise of increasedproductivity, efficiency and savings from running voice over thenetwork. But Yankee Group's Quigley believes the whole area of IPcontent - including video and interactive multimedia - represents asignificant opportunity as well.
"A lot of what has happened in the last year and a half," he pointsout, "has concentrated on services for consumers - MP3s andvideo-on-demand on the Net. But there is also tremendous value inbringing those same types of services to an enterprise, whether it beto have enterprise-wide interactive training sessions or one-to-manyvideo connectivity to off-lay travel expenses."
Video conferencing, or at least multimedia conferencing - voice plusapplication-sharing - should also get a new lease on life from IPconvergence, Fox says.
"We should be trying it more," she argues. "People are tired oftraveling. Nine-eleven should be forcing people to look at this, giventhat [increased airport security] is adding two hours to every tripand a lot of personal grief."
Going Wireless
Our experts agree that CIOs should be investigating wireless datanetworking.
There are two overlapping aspects of this: mobile data networkingusing cellular networks (especially new higher-speed 2.5G networksbased on 1xRTT and EDGE technology), and wirelessly extending theenterprise network using 802.11 broadband fixed wireless technology ina facility or on campus.
One point of overlap is public broadband wireless-access hotspots inhotels, airports, conference centres and coffee shops. Some mobileservice providers are already working on integrating mobile and 802.11hotspot networks to provide customers with seamless, optimal coveragewherever they go.
Angus prefers the expression "mobile business" to the more popular"m-commerce" when referring to this technology's impact on business,because he believes the real opportunities are not in selling thingsto customers over mobile networks but in extending the enterpriseoutside the office. "Where we're seeing quick payoffs is in equippingmobile workers with wireless services," he says.
One of Angus's clients, a company that sells service contracts toconsumers, recently equipped sales people with wirelessly connectedPCs that allow them to log in to a corporate server from the customer'spremises, run quotes based on up-to-the-minute pricing, figurepersonalized variations for prospects, and initiate sales online.
"They tripled their close rate," Angus says. "Before, they would haveto go back to the office, check prices, submit forms. By the time theycame back, half the people had changed their minds. It wasn't a bigapplication - it's not a big company - but it made a huge impact."
Wireless is already on CIO's radar screens, Quigley says. Theyunderstand horizontal applications such as e-mail and Web access. Butmost have not yet grappled with how to apply the technology to theirown operations. They need to do that.
There are significant obstacles, though. The cost of equipping a salesforce of 200 with Pocket PCs or Palms, developing an application,implementing servers and managing a network "is not going to beinsubstantial," Quigley points out. And as Gartner's Fabbi notes,managing wireless devices, many of which have already come into thecompany through the back door because employees bought them to managepersonal information, could become "an incredible nightmare".
"CIOs have a hard enough time managing what's inside their four walls,"Fox notes.
Wirelessly extending the enterprise LAN, allowing employees to stayconnected while they work in the conference room or the foyer, alsomakes increasingly good sense, our commentators say, despite welldocumented but manageable security issues.
Angus points out that the emergence of the multi-service IP networkwill enable another useful wireless service: low-cost wirelesstelephony within the office. If you already have an IP voice systemand an 802.11 wireless LAN, there are vendors with standards-basedwireless phones that will work seamlessly with the phone system overthe wireless network.
Fat Pipes
Some CIOs are already grappling with the question of whether or not tomigrate to Gigabit Ethernet (1 Gbps) network infrastructure- and if so,when? Our experts suggest it should be a less pressing concern formost, although there are exceptions.
Vendors will attempt to force the issue and create demand, Fabbi says,by first introducing network interface cards that offer 10-, 100- and1000-Mbps functionality in one product, and then introducing switcheswith 10-, 100- and 1000-Mbps capabilities. Most CIOs should resistthese inducements, he warns. Increased bandwidth in switching andinterface equipment will generate demand spikes that will createlatency problems for IP voice and put pressure on backboneinfrastructure. And the next level of 10 Gbps backbone infrastructureis still far too expensive, he says.
"If you're in seismic research or medical imaging, or you're a videoproduction house, Gigabit Ethernet may make sense," he says. "Anytimeyou routinely have to move files of 100 MB to 1 GB - wherever time savesmoney or lives. But that represents maybe 10 or 15 per cent of theuser population; the rest of us don't need Gigabit."
Fox adds one more type of candidate to the list: companiescentralizing data storage on a large scale might also need GigabitEthernet in local and wide area networks to support the increasedtraffic that results.
Network Security & Web Enablement
Much of the CIO's concerns around communications now centre aroundhaving the right firewalls and intrusion detection systems in place,and having reliable network infrastructure.
"Clearly security is top of mind and that's probably where it shouldbe," says Gartner's Mark Fabbi. "What do I do in case of terroristattack? We get tons of those kinds of questions. The answers revolvearound building redundancy into network designs, designing datacentres properly, and so on."
Longer term, intrusion detection systems should begin to give way tointrusion prevention systems, according to Gartner. "Are there systemsout there? No, not now. We're just at the brink of a new technologythat will allow us to take the next step in security," Fabbi says.
The current preoccupation with Web enablement - converting enterpriseclient-server applications to work over the Web and making morecorporate content available online - also has medium- and long-termimplications for network infrastructure design.
"We're no longer dealing with just raw packets," Fabbi points out. Tomake Web-enabled applications work well, enterprises must begin toimplement application-layer capabilities in the network to allownetworks to make intelligent use of information in the packet headers.
Delivering content and application services will also mean scaling upnetworks. Scaling up in terms of sheer volume - bandwidth and servercapacity - is straightforward enough, but widespread Web enablementshould also mean scaling up complexity of network design.
For example, rather than simply adding undifferentiated servercapacity, it may now make sense to split content and applications intological chunks (e.g. finance, HR, operations) and assign a differentserver or cluster to each, or add regional servers rather than morecentral server capacity. Then with application-layer capabilitiesdeployed, it becomes possible to streamline routing.
It's still early days, though, for what Fabbi refers to as "intelligentnetwork overlay" technology. While the market for switching androuting equipment is close to $25 billion a year, the market forintelligent overlay technology is only about $500 million.
What About Vendor Viability?
If CIOs aren't worried about whether communications hardware andservices suppliers are viable for the long term, they should be, ourexperts say. The danger signs are everywhere.
The demise of Global Crossings earlier this year was only the latestin a string of bankruptcies among competitive carriers. Canadiancompetitive carrier Group Telecom announced it was seeking bankruptcyprotection in June. WorldCom in the U.S. followed with a similarannouncement. Hardware vendors are also disappearing, or selling orabandoning product lines.
"It's a question a lot of our clients are asking us - Is our suppliergoing to be around next year or next month?" says Ian Angus. "You'dalways get this from customers of Frank's Storm Door & TelephoneCompany, but now we're getting it from people working with the biggestnames in the industry."
No company can afford any longer to have one vendor that supplies itslifeblood, Angus cautions. The supplier that you always relied onbecause they were so good may continue that way, but the headlinesshow that you should not assume it. CIOs should at least be thinkingabout and planning what they will do if that vendor ever stopped beingas good, or disappeared. This goes even for blue-chip communicationscompanies.
CIOs should also be learning more about their vendors. "How many CIOsknow how to read a telecom vendor's annual report?" Fox asks. Theyshould know, she argues. And they should be delving into andunderstanding stock reports too. CIOs also ought to attend vendorconferences themselves rather than sending someone from theirdepartment, Fox adds. Such conferences can be key sources ofintelligence on suppliers.
It goes without saying that the same kind of scrutiny should be paidto prospective new suppliers.
"Price should not be the be all and end all anymore," notes YankeeGroup's Mark Quigley. "Survivability has to play a larger role now."
Well, those are our words of wisdom from the experts. Now it's up toCIOs themselves to make sure the communications ship doesn't sailwithout them.