Netwatcher (ISSN 0890-5800) is a monthly publication of CIMI Corporation. Subscription information is available here . Copyright © 1999, CIMI Corporation. All rights reserved. No publication or reproduction of this document is permitted without the express written consent of CIMI Corporation.

"Buy Cisco" is one answer, but that’s clearly a bit restrictive in terms of buyer interests. In any event, one might wonder whether Cisco is immune to the market conditions, or simply developing the diseases of the others a bit later.

A better answer is to understand what will be required of buyers of LAN products in the future, and how to insure your organization can meet those requirements.

The problem that most buyers have with LANs today is that they are dependent on an equipment vendor to promote a model of LAN evolution, or on an implicit model being broadly accepted in the market. The latter condition is clearly not true; today’s competitive market does not come to consensus easily on any points. The former condition depends on the continued presence of vendor personnel in the account.

We’ve noted in the past that the LAN market was changing in response to both distribution of purchase authority and commoditization in pricing. These changes combine to limit vendors’ abilities to site personnel with a prospect, by limiting the revenue flow from the prospect to both vendor and salesperson. These factors are behind the acquisitions and difficulties in the LAN market, and so buyer problems can be said to relate to the same forces that have caused vendor problems.

One solution to the problem is to adopt a non-vendor-driven model of LAN evolution, something like the switched LAN approaches we’ve outlined in previous issues. This provides a framework into which various products can be integrated and substitutes for active vendor management of evolution.

A key to success in this approach is acceptance of the limits of technology coordination that the company’s own purchasing policies create. For example, if your company lets line departments buy LAN products that cost less than $5,000 without higher-level sign-off, workgroup switches are ceded to independent purchasers and cannot be expected to play a highly specific role in an architecture.

This point has its greatest impact in the ATM LAN space, because procurement of an ATM-ready workgroup switch by a non-centralized, non-policy-driven, buyer is simply not likely to happen. Thus, any ATM LAN strategy would have to be based on the assumption that edge switching would be Ethernet-based, and that any integration of Ethernet and ATM would have to be carried out at the boundary between the workgroup switch and the second-tier network, presumably the place where technical coordination of the LAN purchases would be exercised.

If LAN switch procurement outside the network operations area can contaminate a model-based LAN evolution approach, so can problems with integrating the current equipment. Since the first test a LAN deployment model will face is supporting what’s already deployed, failure here assures quick discrediting of the whole process.

Our research shows that failure here is almost a given, because most LANs have evolved in piece parts based on individual application needs, or by substituting switches on a box-by-box basis for hubs or other dumb concentration devices. Thus, there’s a good chance that the current network doesn’t support any particular model at all.

The first step in facing a LAN problem is to take a stand on LAN equipment deployment. We believe that workgroup switches can be bought at the workgroup or business-buyer level, and that this type of purchasing should not be restricted. The PC revolution showed how useless it is to try to control line organizations’ access to specific technology.

We also believe that it is probably crucial to maintain central control of everything above the workgroup, and also to maintain control of the organizations’ access to WAN services. The linking of workgroups into a network is a technology problem, one that individual business units should never be allowed to solve on their own. Likewise, the use of wide-area services must not be permitted without technical coordination, or there are major security and other risks introduced. It may be possible to create a framework for purchasing higher-level switches or even VPN services by departments (in fact, it is inevitable that this will happen eventually), but the central policy control must remain.

A more controversial issue is the status of basic network software, or even the definition of where "network software" ends and "application software" begins. Letting users select their own NOS opens not only a potential problem in supporting the inevitable diverse selections made, it virtually guarantees that some intra-company communications options will be closed off by incompatibilities. Yet some organizations allow workgroup latitude even at this level.

NOSs should be standardized where possible, except for business units whose internal data repositories are necessarily off-limits to others (payroll comes to mind). Likewise, the decision to create a LAN at the 802.x level or at the IP level should be made company-wide. Finally, the same directory strategy should be adopted across the entire company. We recommend central installation and support of all these elements.

Groupware related to communications, such as e-mail, collaborative software, video multimedia, etc. should also be standardized and controlled at a central policy level, but it may be acceptable to retain distributed purchasing and installation, or even support. The issue here is the extent to which these applications are bound into business practices at a consistent level. If one group uses all these groupware elements integrally, and others use them hardly at all, it may be better to have distributed purchasing and support, and financially fairer as well.

The hardest issue of all to face is the issue of information exchange over the network, and the impact of this on application selection. If everybody picks their own word processor, how useful is an architecture to permit interchange of files? On the other hand, personal productivity tools are a bear to control because the wrong selection can be devastating to an organization with special needs.

One approach that seems to work as well as any is to select a standard for information interchange (Rich Text Format for document files, for example) and allow any application that can both import and export that standard form. In textual information this is pretty easy; in graphics there are many slips that can occur. About the only standard we’ve found there is encapsulated PostScript.

Vendors have helped users over these humps in the past, but in truth their recommendations have been of spotty value. Typically a vendor has a limited product range, and thus knows little about the area in question if it’s outside the comfort zone, or has a full product range and gives self-serving advice because of it. Social consciousness on the part of vendors died about 1992, according to our research, and in most key market areas 100% of buyers believe vendors make substantive misrepresentations on product capabilities.

The topic of voice over IP has probably been developed far beyond the point of active disinterest, from the perspective of most of our readers. We’ve also had the related topic of "convergence". Well, gang, the bad news is that we’re about to have another wave of monologue here. The good news is that IP isn’t a big part of it.

It’s no surprise to anyone that we’re reshaping both the service content and the infrastructure of public networking. It should be no surprise that this process will have some impact on voice services. In that, we agree with the majority view. Where we disagree is in virtually every aspect of how the changes will come about, and what technology will shape them. So, we’re going to take a hard look at the concept of future voice service and architecture.

We’re now in the third year of the Telecom act, and there are many players in every sector of the market that are looking for that mechanical god to start swinging. While there have been some local successes with telecom reform, the truth is that most users of telephony don’t really see any fundamental service difference today. Moreover, we could argue that those differences that do exist were motivated as much by market forces as by regulatory change (cable modems, for example, are probably promoting both ISDN and xDSL).

Are the guys in Washington asleep? Well, yes, of course, but are they more so than usual? Are there any signs they’re waking up? If they are, what impact will the individual strategies of the various players in our complex regulatory framework do to the market at large?

There are signs of big doings. The big overseas telecom giants, long somnambulant in the US market space, suddenly seem bent on re-entry and effective competition here. What do they see? SBC announced a deal with Williams to buy bandwidth to extend its operations out of region. That’s been an option available to SBC without restriction since 1996 (the Act immediately allowed RBOCs to operate outside their regions, it just limited their long-distance options within their region), so why now?

The action by the Supreme Court probably doesn’t signal the end of the foot-dragging, but it does signal the end of the stonewalling. There will still be appeals filed on arcane issues of the Act, but there no longer appears to be an organized position that the Act itself isn’t enforceable by the FCC. Moving the issue out of the state PUCs, where ILECs like the RBOCs have a strong position, makes things harder for the incumbents.

We believe that the Supreme Court’s position is the first step in breaking up the logjam on competitive reform in the industry. It won’t completely resolve the problem of lower Federal Courts’ intervention (8^th Circuit is particularly troublesome in its meddling, and we expect to see more of it), but the Supreme Court seems to be signaling that they’ll move quickly on these matters.

A PUC arbitrator in Texas made that same decision last year, but it was overturned by the PUC on the grounds that the FCC would have to make the final determination on the status of Internet calls, and that in any event a specific interconnect agreement to pay for the calls would supercede any other factors.

The FCC has now made the ruling, and it’s subject to some interpretation. First, the FCC said that it believes current calls are primarily interstate. Second, it indicated (but didn’t say) that it would probably rule in the future that such calls were specifically not subject to reciprocal comp, but rather might be subject to some other yet-to-be-determined measure of compensation. Pending such a rule, the existing state PUC rules and interconnect agreements will prevail.

What this means is that ISPs cannot expect to continue to be compensated for calls that terminate on their modem pools, either directly or when a CLEC acts as a kind of legitimizing agent to evade the fact that ISPs would originate no calls at all, and so the compensation would flow in one direction only. This will probably reduce the movement of ISPs into the CLEC space, and also cut into the profits of CLECs who thought they’d be an agent for multiple ISPs, offering low prices partially justified through the addition of reciprocal comp revenues.

Later this month, the FCC is expected to take an even more dramatic step in a broad rule-making arising out of the RBOCs’ appeals for relief from wholesaling under Section 706 of the Act. The Notice of Proposed Rulemaking issued in September (98-188) indicated that while the FCC had no authority to grant such relief to the ILEC, a CLEC properly spun out of the ILEC and operating under the structural separation requirements of the Act could deploy data-only infrastructure without a wholesale obligation.

Nearly everyone expects that position to stand, and all of the RBOCs are already in some stage of preparing to create such a subsidiary. It would operate not only to provide enhanced (non-voice) services (owning this equipment, so it would be protected from wholesaling), but also probably serve as the out-of-region sales/provisioning arm, and the retail sales arm for all services offered by the ILEC. Thus, the ILEC would become a wholesale-only entity.

This has been the FCC’s strategy all along, in our view. With the ILEC operating as a wholesale player and all residential and business services sold out of a CLEC subsidiary, the ILECs’ compliance with wholesale requirements would be assured by the simple fact that without wholesale elements the CLEC subsidiary would have no infrastructure to exploit with add-on data services like xDSL, or services to sell. The enforcement of the terms of the Act then becomes nothing more than guaranteeing that the other CLEC players get the same deal as the CLEC/ILEC subsidiary.

Most of what the politicians want to promise us isn’t theirs to give. A future of superfast home Internet connectivity demands a subsidy source to the ISPs, which demands a mechanism for limiting the Internet’s traffic to things people are willing to subsidize. Given the industry’s counter-culture fervor, that’s unlikely in the near term.

Congress probably won’t intervene in the reform process this year unless things go seriously off-track. We don’t think that will happen based on recent activity. Thus, we think the industry is safe from further regulatory huff-and-puff—for now.

Access the index of CIMI Corporation's recent newsletters.