Tech Savvy
About Kumar, Brocade and the RealMoney.com Passion Pit
By Jim Seymour
Special to TheStreet.com
8/22/00 7:43 PM ET
URL: thestreet.com
Interesting hoo-haw Tuesday morning on the RealMoney.com Columnist Conversation board over the prospects for Brocade (BRCD:Nasdaq). (If you're not following the Columnist Conversations, you're missing a lot of the value of RealMoney.com. Go back through Tuesday's posts and you'll see what I mean.)
Analyst Ashok Kumar, with Piper Jaffray, released a research report this morning casting doubt on both the future of Brocade, and on the sustainability of its current sky-high share price.
Kumar questioned the future of the fibre-channel switching market -- more on fibre-channel technology in a minute -- and worried that a stock with a current price-to-earnings ratio of 569 (yes, that's correct: five hundred and sixty nine!) has no place to go but down.
Jim Cramer loves Brocade, as he has written here; he weighed in. And other RealMoney.com columnists have their own perspectives on the company, and on Kumar. Herb Greenberg recalls when Kumar, then based in Minneapolis, praised local stalwart Ancor Communications (ANCR:Nasdaq) -- a company Herb has long thought has done its shareholders wrong. He also remembers Kumar "touting Ancor as the second coming of Brocade," and now sees Kumar's comments on Brocade as proof that Kumar is painting the whole fibre-channel switching area with the same tar brush.
Adam Lashinsky worries about Kumar's "flip-flops," and sees in Kumar's former support for Ancor more than a trace of the "homer" -- sportswriter talk for a guy who constantly puffs up the accomplishments and prospects of his paper's hometown teams and players.
Me? I think Kumar nailed it.
He made two very good points. One's easy; the other, less so.
The easy one: With a price-to-earnings ratio like Brocade's, you've got to be very careful. No matter how much you believe in momentum trading, buying a stock with a 500-plus P/E is risky. Period.
Worse, what's a nosebleed-P/E like that based on?
The only reasonable answer: sky-high expectations for future earnings.
Which leads directly to Kumar's second point: Fibre channel has a limited, and diminishing, future.
We have to backtrack a little to make sense of this. Fibre-channel communications protocols and systems were developed as a means of moving data among and between Storage Area Networks (SAN) systems and their connected servers.
The SAN idea arose when computer-network architects pointed out that it wasn't necessarily smart to mix the server and storage functions. With demands for storage capacity growing, and with the increasing need to serve that data up to very large numbers of users who might be widely dispersed geographically, why not put together banks of high-speed data-storage devices, not installed inside the tower cases housing a networks' servers but on their own ... and let that data be managed in ways never possible in the complex, multifunction world of the computer server?
Good idea. It was necessary, then, to find a very fast means of passing that data back and forth between these banks of storage devices, and back and forth to the widely dispersed servers through which requests were made for the data. Enter fibre channel, which EMC (EMC:NYSE) and IBM (IBM:NYSE) have long been touting as the right answer.
In many ways, fibre channel is an updated form of the older Small Computer Systems Interface (SCSI) system we've long used to connect hard disks, scanners and other devices to PCs. It is fast -- up to 850 megabits per second. And so Brocade emerged, as the leading producer of fibre-channel "mesh" switches for SANs.
But more recently a better idea has come along: With so many companies now having very fast backbones in place, why not just switch over to standard Internet Protocol communications, instead of fibre-channel systems, for that work? IP is well understood and well proven; it is ubiquitous; it does not require a company to, in effect, start over with new tech standards and new tech people.
Standard Ethernet running IP protocols, was too slow for the job. But we now have Gigabit Ethernet in place, with 1,000 megabits per second speed. That's not just fast -- it's faster than fibre channel.
And it uses standard equipment and has much lower costs than fibre-channel implementations.
Pick up any information technology industry trade paper, go to any IT trade show over the past six months, and you'll find a lot of buzz on using Gigabit Ethernet for SANs. And hardly any on fibre channel.
Part of the confusion here grows from the continuing success of the SAN idea. Sure, the SAN business is exploding. Sure, it's a great idea. Sure, lots of companies are moving to SANs in a big way. But not all with fibre channel. And over the next two or three years, I think Gigabit Ethernet is going to muscle fibre channel to the sidelines in SANs.
There are a couple of technical issues here we have to discuss. Fibre-channel boosters, especially Brocade, love to slam the idea of using Gigabit Ethernet to manage SANs because of the "overhead" of TCP/IP communications, and because of the "out-of-order" problem.
Both refer to the computer busywork required in IP communications to tag, track, resend and reorder data packets. Those who read my column here a couple of weeks ago on packet-switching vs. circuit-switching will recall that in shipping packets around, via IP, the packets sometimes get lost, and often arrive out of sequence. The chip managing the process at the receiving end asks for resends of missing or garbled packets, and as all packets arrive, it must check that they're in the right sequence -- and when they're not, reorder them.
This takes a certain amount of computing-power overhead and necessarily means net transmission speeds are somewhat lower. (On the other hand, remember that this is done in the name of data integrity, so it's hardly a bad thing.)
The fibre-channel boosters love to claim that overhead makes IP transmissions, at any speed, unsuitable for SANs where performance is vital. They love to snow people with the presumed inevitability of a fibre-channel takeover.
Yes, but, but. ... There are other issues. Like standards. Like economics. Like the way the world is moving.
I have no case to make for Ashok Kumar. I have often disagreed with him on his calls, and I will again, I'm sure, in the future. But not this time: He nailed it on Brocade.
One last thought: If I think (as I do) that Brocade can probably keep growing a little longer, no matter how insane the P/E gets, before the world catches on, and I worry only that in a year or two or three their fibre-channel business will start to recede ... so what?
Isn't investing about making money now, here in JJC's Church of What's Happenin' Right Now?
Yes, if you're a trader. Play the Brocade game a little longer -- the company is flying right now -- but be careful. But for investors? For people who want a little security with their capital appreciation?
Even if Brocade is the famous Last Man Standing as the fibre-channel business winds down, that still isn't going to be an enviable position. A company like Brocade -- and especially, a price like Brocade's -- demands a nonstop growth curve. And that curve is coming to an end for Brocade.
------------------------------------------------------------------------ |
