Friday, December 18, 2015

Connecting Clouds

A few weeks ago, I posted "Cloudy weather ahead" citing HP's abandonment of the Public Cloud.  This was days before the HPE/HPI separation, and the reaction of many to the post was 'confusion' to say the least.

"How could HP be abandoning the Cloud?" was the most popular refrain.

And of course they weren't, in terms of hybrid clouds, etc.

But this cloud business is ... indeed cloudy.

In today's SF Chronicle, I published a small article entitled, "Last mile: the home stretch for Hybrid Clouds"   You can see it at http://goo.gl/1JLunW

The observation essentially is that where the Cloud technology and deployment issues stand is akin to where connecting networks was in about 1985, before widespread adoption and deployment of routers.

Having done some work re Cisco history, I was struck by how few firms understood the power and the ease of connection for multiple disparate networks.  It took nearly a decade for most of the Fortune 500 to 'get on board,' astonishing as that might seem today.

IT directors, recall, are not paid to be risk-takers, but instead to be fully risk-averse.  We med with many of them when I ran R&D/Product Marketing for Informix circa 1991.  After one dinner with a number of IT CIOs, my wife on the way home said, "WHO ARE THESE PEOPLE?  DULLER THAN...."

It fits, and still fits today.  The issues today of course are complicated by cyber-security issues, contending vendors, and costly revamping of the core stack privileges.  But they were back then too....

Anyway, I hope you enjoy the article

Silicon Photonics

One candidate for considerably improved computing performance is SILICON PHOTONICS

To the question of what has DARPA or other government agencies done for us lately, the news from the Obama administration on this front seems encouraging.  To wit, the formulation of AIM Photonics last year.  Michael Liehr, at SUNY, is the CEO/Director.  I don't know him.  The Deputy director is John Bowers, an incredibly capable researcher based at UCSB (Univ of Calif at Santa Barbara).  He is assisted by Rod Alferness, also from UCSB.I;ve worked with them in the past, and have the highest regard for them.
   

The task though is a tough one.  It is housed at SUNY Albany, with strong goals and capable leaders.  But recall SEMATECH, and its abortive start in the mid-1980s, even though DARPA and Craig Fields did monumental work to aid U.S. firms in their fight vis-a-vis Asian chip manufacturers.
SEMATECH eventually morphed, and moved to SUNY Albany, creating the NanoTech Complex, a key reason that the Si Photonics labs are being located there as well.

Also recall MCC in Austin, again with much fanfare (Dields eventually ran it) to 'save American leadership in software'  

Let's hope the lessons of those approaches are embodied in the AIM Photonics effort.

Below is the first slide of a quasi-public document describing the new facility




There are two levels of support beyond the government sponsorship-- Industry with three tiers of supporting level, and academia with an analogous three tiers.   Below are the logos of the first tier of industry--clearly some big names (recall, for example, that Intel stood aside from SEMATECH until Bob Noyce was willing to head it up).  Keysight (our old instrument friends from HP) are in Tier 3.  Cisco, Juniper and Texas Inst have all said "we'll play" but haven't ante'd as yet, so are in no Tier.




Herewith the lead schools  (note, Caltech and Stanford are in Tier 3, Berkeley and UC San Diego in Tier 2).



I'd be interested in any thoughts you have on this topic

Tuesday, December 15, 2015

Looking forward

We're well into "the new HP era", right?   Six weeks in, in fact.  And the word on "the machine" is muted, but worth discussioin.  As are a lot of other "what now" topics.

One that hasn't gotten much mention from HP in awhile is Quantum Computing, a worthy contender (for many theoretical years) to HP's approach.

Last week, a gaggle of Google researchers announced some interesting results from their analysis with a D-wave machine--including a result that sped up operations by some 100 million times.  Wow!  We should get one of those machines.

And then, an MIT researcher, Scott Aaronson, gave an ACM interview re 'what does this all mean?'  You might find it interesting reading at: http://cacm.acm.org/careers/195454-scott-aaronson-on-googles-new-quantum-computing-paper/fulltext

Net net--we're all still waiting.