Acme Corp. Network Assessment Recommendation Digital Equipment Corp. takes a look at the networking needs of an insurance company and makes recommendations. In Word for Windows 6.0 format. The Networking Dilemma White paper by Digital and Strategic Networks Consulting that discusses how to identify and deal with networking costs. Strategic Networks Consulting's online publications EnterprisePro demo Network perfmon tool from International Network Services. comp. dcom.net- management archive Also has net-management links. Csenger is a freelance writer based in the Chicago area. He can be contacted at mcsenger@ mcs.net.

By Michael Csenger
Network World, 7/7/97

It seems surprising, but most companies do not - cannot - lay plans for technology to support new business applications being rolled out on their networks. Conducting a needs analysis to determine what networking resources are on hand, how they affect each other and what might need to be changed or added to support new goals falls outside the business processes of most organizations.

Proper needs assessment would spare the mess and wasted assets that result when projects fall apart. But it doesn't happen often. Instead, there's a revolving scenario that defies the team-management principles with which modern business practice is supposedly imbued. New applications and entire new business plans are piled onto a network until it breaks. Beepers go off and the smoke rises - the network is down, it has to be fixed, some money gets thrown at it and the next round begins.

There's the case of an IBM mainframe shop that rolled out a document imaging system over a NetWare LAN for $250,000, or so it thought. Several unexpected server upgrades, backbone expansions and WAN bandwidth boosts later, the implementation cost an additional $340,000. Factor in considerable network downtime, and it's hard to account for such an obvious planning blunder.

There is a better way, but it's not simple, it costs more up front than most operating budgets allow, and its success requires the efforts of a truly unified team. Conducting a needs assessment is an acid test for networking professionals. It's ultimately a business re-engineering task that convinces senior management and other department heads to come to grips with networking's role in the enterprise.

Although regarded as a vital asset, networking is still budgeted as a back-office expense. It's expected to perform and improve upon itself through wonders of technology that boards of directors read about in magazines. ''They go to a party and hear about frame relay, and that's all they know,'' says John McClelland, senior network consultant with Digital Equipment Corp.'s Operations Management Services in Colorado Springs. ''People in business do not understand the technology, so we see bandwidth increases of 30% to 50% being met with budget increases of 3% to 5%.''

Even companies that have the technology in hand aren't mastering it well, says Michael Kennedy, director of consulting services at Strategic Networks Consulting, Inc., in Rockland, Mass. For example, sites that employ robust network management tools often lack accompanying business processes for carefully managing capacity expansion. ''There's a lot of running around instead with screwdrivers to fix the squeaky wheel,'' Kennedy says.

Network consultants, integrators and the more candid net managers all echo the same theme: real needs assessment doesn't happen because network planning is conducted under heavy fire when the infrastructure fails. ''That's about 80% of our business,'' says Alan Weingarten, a network consultant with Digital's Digital Services Division in Albuquerque, N.M. ''Companies call us in to help figure out why their network just broke.''

Unless they're embarking on a technology migration large enough to grab everyone's attention, companies rarely assess their requirements and plan ahead. Networks can be healed, but entire projects fail and die as a result of this poor planning.

Each undertaking is as different as the business factors driving it, but following some basic steps can get you through the needs assessment process.

Step 1: Attitude adjustment

The real labor consists of establishing the lines of communication and determining what the project needs to accomplish. Actually figuring out the technology requirements for new applications will be grunt work in comparison. The bits and bytes can be delegated - shepherding the overall process is the real crux.

The people running large, well-oiled shops inhale sharply when asked for general guidelines on how to conduct a needs assessment.

''There are no simple rules of thumb that you can boil this down to,'' says Jerry Krause, director of network services for United Airline, Inc.'s Apollo Travel Services. ''There are whole books written on the subject of developing business proposals. It's a matter of keeping communications clear and open at every level and at every phase of the project.''

Attitude adjustment means accepting the fact that even though you're the network manager, you probably will not gain control of how the network is used. This dawning recognition has already led to a strategy of overbuilding campus networks whenever possible to accommodate at least the next round of traffic.

Tom Nolle, president of CIMI Corp., a consultancy in Voorhees, N.J., recently led a seminar during which he asked a room full of network managers if they had any control whatsoever over the applications being run across their networks. ''These are large organizations with enough forethought to at least send their people to seminars, yet the response was less than one out of a hundred,'' Nolle say.

One manager of a $1 million-plus network budget admits, ''I don't even know what's running over my network. All I see is more traffic, and I try to stay on top of the curve.''

LAN switching is welcomed because it's cheap enough per port to make reserve capacity more affordable, Nolle says. ''Networking people have now realized that the business people really will go ahead with a project that breaks the network. So they're building forward just as far as cost-justification can allow, enough to absorb that next project and prepare a new business case when the reserves are drained.''

Step 2: Open dialogue

Be prepared to explain the potential benefits and return on investment to senior management to get a needs assessment off the ground. A full-blown consulting contract that starts with strategic planning and stops just short of a request for proposal can range from $150,000 to $500,000. More affordable out-tasking relations that help with key parts of the process start at $10,000 to $20,000 but will rise to more than six figures for larger networks. Work conducted entirely in-house will cost less because profit won't be factored into the figures.

''In senior management's mind, you're going to spend a lot of money on the network and not be able to do anything more with it than what they already see,'' Krause says. ''That's the disconnect: In their minds, you still have all the same applications, the same PCs, so they see nothing you've done to improve the business. You have to convince the CEO that there is, in fact, a payback - that you're creating the potential for new capabilities or improving current capabilities so there'll be less downtime. You present a business case and speak purely in those terms.''

''The most broken environments we've seen have very little communication going on between the technology people and the business people, which is why they're in the mess they're in,'' McClelland says. Needs assessment will often arise in an atmosphere of stress and denial, and this finger-pointing makes it that much more difficult to establish the dialogue that must begin.

McClelland offers some pointers. ''You want to steer the conversation to where there's no blame being passed around. If you're in a meeting with the conflicting parties who triggered the problem, you don't ever want to push them to the point where they feel they have to defend what they've done in the past,'' he says. ''I don't care how they've gotten to this point - we need to talk now about where to go.''

Step 3: Determine user requirements

Here, too, it gets political. ''Getting people to work through these basic requirements and end-user goals is business process re-engineering, and you have to treat it as such,'' says Glenn James, a partner with Deloitte & Touche Consulting Group, in Atlanta. ''The technology can be beat to fit the people better than the people can be beat to fit the technology.''

That's a mantra that technologists easily lose among the details. People at the network level see their work in terms of IP and IPX traffic, not in terms of how much traffic is stemming from specific applications or processes, Kennedy says. Baselining tools exist for this sort of analysis, but they cough up enormous volumes of data.

Step 4: Baselining

Baselining is the first hands-on step of a needs assessment. It can be done in three ways - through a consultant, in-house using a turnkey solution or outsourced to a provider of baselining services. The key aspect is not to just inventory what's on the net, but to figure out exactly how different applications are consuming resources. This can be done in surprising detail, but not without a steep learning curve. Here's where outside help comes in.

First, network consultants and out-taskers can help break a communications impasse. They provide the baselining insight to offer a clear picture of network usage and trends. Finally, they serve the standard consulting role of helping define business goals and translate them into technical requirements and a viable business plan.

Baselining isn't cheap, but it's essential for capacity planning. Baselining tools use Remote Monitoring 2 probes to gather Layer 3 data from devices across the network and require considerable expertise to implement and administer. Sites can run turnkey solutions in-house, but will usually need a full-time administrator to update changes to the network over time. A baseline must be run over a period of at least three months to provide real trending data. The baseline offers even greater value when used as a full-time planning tool wedded to the overall network management system.

Concord Communications, Inc. of Marlborough, Mass., sells a turnkey package called Network Health, which can factor the cost basis for different applications traffic. Concord and other vendors also partner with service providers and consultants who use the tools as part of their service package. International Network Services (INS) provides Enterprise Pro (EPro), a Web-based tool that customers can use on their own or outsource through INS. Building on its own experience with turnkey solutions, INS crafted the out-tasking service as a simpler alternative.

A baselining service such as EPro costs roughly $2,000 to $3,000 per month for a mid-size corporation, and can run as high as $10,000, depending on network complexity. Concord's platform ranges from $10,000 to $30,000.

Whatever the choice, vendors, consultants and service providers point out that these tools are complex and don't yield worthwhile results without considerable user training. Qualified feedback is essential.

Step 5: Nuts and bolts

Not to oversimplify, but this phase of the project is probably the easiest because it builds on the previous four and gets the network planner back on familiar technical ground.

''If you're trying to do something based on rational planning, when you arrive at this point, you'll be able to do it successfully,'' Nolle says. ''Getting the skills to move forward is much less complicated than figuring out where you want to go.''

Other consultants agree but warn that the job is still far from over. ''Customers usually understand the limits of their database engines and applications servers,'' McClelland says. ''What they don't understand is the impact on the strings that tie it all together - there are a lot more systems analysts out there than network analysts.''

There is no codified plan for figuring out specific needs and implementing their execution, McClelland says. ''It's a matter of common sense and flexibility. You need to be astute enough to see problems happening and then have the knowledge to figure out why.''

Much of the planning at this point happens on whiteboards and paper. Network modeling tools can help, but, like baselining tools, these require considerable expertise to be used effectively. ''You're better off doing it by hand,'' McClelland says, although others disagree, saying that the ability to model network permutations and different traffic characteristics is key to understanding what future problems to expect and how much they'll cost to fix. Step 6: Go the distance

This is often the most important aspect of planning, according to Nolle. ''The application has to be trialed by the people who are going to use it - nothing else will cement the connection between users and their technology.''

This need for hands-on feedback is the main reason why shrink-wrapped software and low-performance database programs are so often pressed into service. ''In the '80s and early '90s, users got fed up with big, custom projects because they had no idea how the solution would fit them until a massive part of the undertaking was already done,'' Nolle says. ''People couldn't communicate their needs effectively enough to get projects launched in the right direction. You still have to guard against that happening.''

Step 7: Put it in writing

''It's time for the dreaded 'Thou Shalt Not' document,'' Weingarten says. ''You need to establish centralized procedures for basic things like administering IP addresses. It doesn't have to be all that rigid, - you just need to document basic security and administrative procedures, or you'll be starting all over again before you know it. With switches today, you really want to keep gurus down the hall from hooking up their neighbors, and that means you'd better meet your obligation to support those needs.''

Step 8: Better days ahead

But the vast majority of networking shops are just coming to grips with frame relay, and future trends can be as confusing as always. Compounding the technical environment are business factors such as mergers, acquisitions, and continued downsizing. Today's warped disregard for technology planning stems in part from the ensuing confusion, no doubt, which ought to drive home one salient point: Net managers who speak technology and business well enough to plead the case for a needs assessment will arrive at healthier surroundings than those who don't.