Bay tries another route around Cisco

With all that's been happening of late - AT&T buying TCI, fires in Florida, floods in the Northeast, and probably locusts somewhere - most people overlooked a relatively quiet announcement from Bay Networks, Inc. last week.

Bay, which has suffered more than most as a result of Cisco Systems, Inc.'s routing dominance - although partial blame for the suffering has to be dealt to Bay's management team as well - has adopted a different strategy for promoting its routing products: open code licensing.

Bay recently created the Bay Open Routing Division to license a version of the company's routing code known as Bay Routing Services (BayRS). On the face of it, this seems like a reasonable thing to do, since although Bay has not kept up with Cisco in the router space, it has been developing some really nice routing code. But in typical Bay fashion, it instead has chosen to rely upon the code of another firm, Phase2 Networks. Phase2, which was acquired last week by Bay, is presently providing its portable routing code to over "50 leading network equipment vendors, consisting of both Fortune 500 and innovative start-up companies."

The theory behind the acquisition is that Bay should be able to blend Phase2's low-end routing code with Bay's high-end routing code and produce the ultimate third-party routing software that will dominate the part of the industry (about 30%) that Cisco does not already control. While Bay probably won't do that anytime soon, it does stand a good chance to improve its position in the routing code market.

Just what is this market? To understand it, you must recognize that the core of most routing engines are relatively simple and relatively similar. Routing itself is a basic technology. It is only when we begin to add "vendor value-add" that it becomes extremely complicated and multivendor interoperability goes out the window.

However, while the basic core is well understood, it is not necessarily easy, or cost effective, for many firms to write their own code. Thus, firms such as Phase2 have been filling a small market niche for either major firms looking to break into the routing market (and that realize they don't have the internal expertise/time to develop their own routing code) and emerging start-ups that recognize that they, too, lack either the funding resources or the manpower or time to start from scratch. In fact, very few of the emerging gigabit switching/routing vendors have developed their own routing code from scratch. It has purchased the basic core, or sometimes the entire package, from other vendors.

This has been going on quite a while. Remember OpenROUTE Networks (a la Proteon)? They have been providing core routing code to IBM and Digital for years. Cisco and Wellfleet also dabbled in this market, providing core routing code to "non-Internetworking" companies looking to integrate routing into their legacy network products.

For Bay, this could turn out to be a good move. Anything the company can do to strengthen its position in the routing market is a positive step - especially if it leads to improved interoperability between Bay and many of the emerging start-ups.

However, it should be pointed out that no vendor has been able to establish itself as a leader in routing through the licensing of code - even Cisco has been less than successful in its attempt to "IOS-ize" the world.

The people who should really be concerned about this are end users. As we continue to see a shift back toward intelligent networking and the use of high-performance routing code at the center of the network, expect that routing will play a core role in many emerging companies.

However, routing may not be a core competency of the firm's research and development organization. Issues such as how long it takes to troubleshoot and fix software glitches could become very important. So while it may be in the best interests of a start-up to buy its core code from Bay, it may not be in the user's best interests if problems with the code implementation occur.

Given that Bay is a relatively stable company -- in a manner of speaking, we still have yet to see the completion of the Nortel deal and it is not clear how that integration will impact this new work cycle -- the risks of bad code should be extremely low. However, the risks of bad third-party implementations are likely to be somewhat higher.

Before users purchase products involving routing, they should closely query vendors regarding where they obtained their source code and how that code is maintained. Unfortunately, not all code purchasers are created equal, and some are likely to get better responsiveness from Bay's internal engineers than others. Users should ask for written "problem escalation policies" before accepting a vendor's word that the vendor and Bay are "real tight - one of our engineers used to work at Bay." Further, the quality of the code integration and the value-add placed there by the purchasing vendor is anybody's guess. So users should be very careful about understanding just how much, if any, of Bay's code has been enhanced or how much "value code" has been added on top of Bay's product.

With this in mind, users should also understand that this opens up the possibility for a larger wave of start-up vendors that know hardware well enough to leverage Bay's extensive experience in routing (along with Phase2) and build some really hot products, relatively fast. But tread carefully, as the path to routing nirvana is anything but smooth.