SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Strategies & Market Trends : Gorilla and King Portfolio Candidates -- Ignore unavailable to you. Want to Upgrade?


To: Apollo who wrote (5407)8/25/1999 12:34:00 PM
From: Uncle Frank  Read Replies (1) | Respond to of 54805
 
You're absolutely correct, Apollo; the comment about fab costs should not have been made with respect to rmbs. Their revenue will come through licensing their ipr.

Thanks for correcting my muddled analysis.

Frank



To: Apollo who wrote (5407)8/25/1999 1:00:00 PM
From: Apollo  Respond to of 54805
 
Sounds like more good news for Intel, as it presses upstream into higher margin markets......

According to Cnet: Future I/O and NGIO set to merge specs.
If true, this will be a very good thing. From what I understand, if the two standards don't converge, it would probably get pretty bloody.

news.com

Truce in the server technology wars
By Stephen Shankland
Staff Writer, CNET News.com
August 24, 1999, 1:10 p.m. PT
A battle of the titans over a crucial server computer technology standard could end in the next few days.

The battle pitted Intel against three of its biggest chip customers, Hewlett-Packard, IBM, and Compaq. At stake has been the future design of servers--the high-speed, high-end computers that are the brains behind the Internet and most other networks.

The Intel camp, along with Key allies including Dell and Sun Microsystems, favored a standard called Next-Generation Input/Output, or NGIO. HP, IBM, and Compaq initiated Future I/O, and rounded up support from 3Com, Adaptec, and Cisco. The specifications govern how equipment such as network cards or disk systems plug into the servers.

"The two specs are going to merge," said a source familiar with the negotiations today. The companies involved will choose a new name for the standard and plan to have the technology ready in time for "McKinley," the second in Intel's line of 64-bit chips.

A merger would lift a huge burden of uncertainty from all the companies involved. Moving into the high end of the market is a key part of Intel's plans, but it would be much more difficult without the support of HP, Compaq, and IBM, all of whom already sell their own high-end proprietary systems.

And for the server makers, a single standard brings the advantages of the Intel business model, where standardization means components ship in higher volumes and therefore cost less.

The lead negotiators from the two sides, who have been meeting for months, have agreed on a proposal and will take it back to their respective camps so each side can vote, sources familiar with the issue said. A Future I/O Alliance vote could happen as soon as this week, but one source said Intel had hoped to delay the vote a little, likely so the company could announce the agreement at the Intel Developer Forum that begins August 31.

Members of the Future I/O Alliance will vote in favor of the merger plan, an industry source said.

The dispute could have been almost as big as they come for the computing industry, said Aberdeen Group analyst James Gruener. "It had the potential to be seven or eight on a scale of one to ten," he said.

The biggest winners in a merged standard will be server customers, who will have one less technology choice they have to worry about; peripheral card manufacturers, who will have a simpler time designing products; and the server companies, who will have an easier time selling their products, Gruener said.

Both sides have been easing toward compromise for months but differences persisted. The Future I/O proponents argued that NGIO couldn't transfer data fast enough and therefore didn't merit the difficulties of a system design change. The NGIO camp said that Future I/O would arrive late and was too expensive, and that they were working on a faster, "fat pipes" version of NGIO to deal with the data transfer speed issue. Fat pipes refers to a faster way to transfer data from servers to components such as network cards.

The proposal resolves these technical concerns, as well as issues of how the companies will govern the standard and how they will handle the thorny issue of sharing intellectual property.

Sun Microsystems, which makes its own high-end servers and has been leading the fat pipes NGIO development, had pushed for an August deadline to merge the specifications, sources said.

Intel is in the driver's seat when it comes to the future server design, said Dataquest analyst Kimball Brown. Though the Future I/O camp had hoped to stall the process, the Intel camp was able to show it was moving ahead, he said.

Version 1.0 of NGIO was released in July, an event that provoked criticism from HP's Martin Whittaker of the Future I/O camp. Whittaker said the NGIO advocates said a revised version would be coming later, in essence admitting that the first version wasn't good enough.

In an interview earlier this month, a Sun executive, Charles Andres, declined to comment on the deadline but said: "To be pragmatic about this, if you're going to get a merge, it's better to try to get a merge to happen sooner rather than later. It gets more difficult as time goes on."

Andres said the Future I/O and the fat pipes version of NGIO were where the specifications would be merged.

"It very well may be the features and benefits in the Future I/O specification...would marry very nicely" with NGIO fat pipes, he said. Both use a similar number of wires connecting into and out of the input/output subsystem, he said.

"From a technical standpoint, a merge is very possible. We could stand to get a specification that has a lot of the good features of both," Andres said.

Representatives from NGIO declined to comment on specifics of negotiations, reiterating the position that they would like to see a single specification.