Eco Sun rising with the UltraSPARC T1

Eco Sun rising with the UltraSPARC T1

Summary: In his recent blog titled "Recipe for Winning Chip Battles," Sun's Jonathan Schwartz makes a case for his company's ascendency based on having a high volume operating system (Solaris, with reportedly more than 3.2 million downloads since it went open source, mostly on non-Sun hardware) and what he calls the fastest chip on earth, the soon to be released, power sipping UltraSPARC T1 (formerly known as Niagara).

TOPICS: Oracle

niagara.jpgIn his recent blog titled "Recipe for Winning Chip Battles," Sun's Jonathan Schwartz makes a case for his company's ascendency based on having a high volume operating system (Solaris, with reportedly more than 3.2 million downloads since it went open source, mostly on non-Sun hardware) and what he calls the fastest chip on earth, the soon to be released, power sipping UltraSPARC T1 (formerly known as Niagara). Schwartz understands the market dynamics (Windows on x86, for example) that get Sun to more volume. He writes:

The chip is built for Solaris, and built for the internet - for jobs like searching, web serving, video streaming or ripping through database transactions (we will be admittedly weak for bomb simulation). Niagara will run existing Solaris apps without recompilation - offering complete binary compatibility for the massive SPARC installed base, and a simple recompile to move from x64 (or vice versa). There will be nothing like it in the industry - and its arrival will give us the two strongest industry standard server lineups [Niagara and AMD Opteron servers] the market's ever seen.

Schwartz describes the new SPARC processor as taking the concept of dual core processors to an "absolute extreme," with up to eight cores, each capable of running four threads simultaneously.

"Doing the math, we'll be delivering a 32-way chip, running 9.6GHz, which sips power (about 70 watts). On performance-per-watt metrics, we believe we'll be a factor of 5 better than what IBM just announced," Schwartz wrote. Sun's PR materials also point to being the first processor with four memory controllers, more security, better crypto processing performance.

Of course, Schwartz is waving the Sun flag and taunting competitors with his 'fastest processor' and  Solaris download count. Now it's a question of whether the combination will get Sun back in gear, along with the new Galaxy Opteron x64 servers. Systems will be available before the end of the year, as well as more independent testing to validate the fastest chip claims. 

If Sun really wants to soak up the server volume, what about making Linux work with the T1? Or, is this an all Sun/Solaris ecosystem play? "We've been approached by many folks to run Linux on T1-- and we are aggressive supporters of running more open source, beyond Solaris/OpenSolaris," Schwartz said in an email. "Nothing would make me happier."

However, the eco-performance of the T1 may turn out to be the major differentiator. Sun claims that the T1 is the world's "first eco-responsible processor." It's Sun's latest marketing theme, but it's a focus that should be at the top of any manufacturer's list of goals.

At Sun's Summit on 21st Century Eco-Responsibility today, CEO Scott McNealy said that Sun was going to make money in a socially and eco-responsible way. Conveniently, the power consumption characteristics of the T1 are impressive. Datacenters owners are tormented by heat generation and power consumption costs. Based on Sun's calculations, if half of the entry-level servers sold in the last three years were replaced with T1s, the equivalent of the carbon dioxide emissions produced by 1,000,000 SUVs would be eliminated each year. And, Sun is imaging the power savings it could deliver to Google and other major Web destinations and high volume transaction operations, and the revenue it will generate.

Sun has made a lot of bets that are somewhat ahead of the market--$1 per CPU compute utility, promoting grids, software subscriptions for an enterprise and desktop stack, open sourcing its software and thin clients. The UltraSPARC T1 is another bet.

The bets all have a common foundation--commoditizing infrastructure. The T1 or enterprise software stack isn't a commodity, but the computing cycles and services delivered are moving in that direction. The T1 (what does the 'T' stand for? Terminator?) may be the piece that helps connect all of Sun's efforts around creating a clear value proposition that is measurably different from competitors.

Update: The "T" in UltraSPARC T1 stand for "Throughput."  

Topic: Oracle

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Sun is the next Digital ....

    Jonathan is correct in that "volume begets value", however Sun is not a volume company. In 2001 we at META Group wrote an article called "Sun is the next digital unless?" on the basis that server infrastructure had shifted to a commodity market where volume is king and Suns traditional technology leadership in SPAC and Solaris where of little value.

    * In CPU architecture x32/64 is the volume play, not SPARC. Sun was very late to this party and brings minimal value. Their AMD systems are very impressive but easily copied by the other vendors.

    * In the OS space Windows and Linux are the volume play, not Solaris. Sure Solaris is superior, but wasn?t MVS in its day! By failing to adequately support x86 though the '90s Sun failed to establish Solaris as the alternative volume OS. Now with two strong leaders, Sun's chances of capturing this are slim.

    * In the web server market ISS and Apache are the volume leaders, not iPlanet/Sun ONE. In the app server space .NET and WebSphere & JBOSS are the volume plays, not Sun. In '00 Sun failed to leverage its then strong Solaris brand to drive volume in either of these spaces.

    Over the 5 years that I covered Sun as a META Group analyst not once did they get it right in the volume market! Sun has been a classic victim of ?disruptive technology? (x86 and Linux) in the sense of Clayton M. Christensen?s ?The Innovators Dilemma? ( While open sourcing Solaris and building a brilliant new CPU architecture are interesting, I don?t see this halting Sun?s long slide into irrelevance.
    • Sun volume

      Kevin, No chance that the T1 could be a volume play for big datacenters given the specs...or is too far of leap to catch x86/Linux multicores...and Intel catch up on the performance/watts maybe?
  • SUN Now Above and Beyond

    The UltraSPARC T1 and upcoming (1U High) Sun Fire T1000 and (2U High) Sun Fire T2000 servers are going to change everything in the industry.

    With being able to run 32 threads in a 1U or 2U high form factor - there is no possible business reason why someone could justify any other Tier 0 Web Server infrastructure!

    This will wipe out any cost AND performance advantage for web services by "scaling out" multiple blades or servers from competing vendors. The royalties for firmware, hypervisor, and OS alone will make them impractical in comparison.

    With Solaris 10 and Containers, it makes infrastructure management from a TCO standpoint absolutely phenominal!

    Any business can place their application in a container, when they need more resources, drag-and-drop (via N1) their application container from an 4 Core 16 Threaded server to an 8 Core 32 Threaded server and experience a painless upgrade.

    From a D/R perspective, place the applications in an N1 Container and have any help-desk agent get the signal to fail over a critical application, drag & drop the application from the (failed) production server to a (shared) disaster recovery server and have people re-connect their applications for a tiny hardware cost.

    No one in the industry offers the Hardware, OS, and Application Support IT infrastructure that SUN offers. SUN FINALLY has a story to sell!

    If every CFO understood the cost savings of going with SUN, every CTO & CIO understood the technological benefits... there would be very few competitors left in the market place - because there are no price, performance, or out-of-the-box infrastructure management competitors. There has been very little innovation in the industry (with the exception of dual core CPU) for the past 6 years... this is a nice change.

    SUN had taken 6 years to re-build itself, but the technology is 5 years ahead of any competitor... and this translates into hundreds of thousands of bottom line dollars to any business.

    If a data center based on these platforms and SUN went out of business the next day, the investment would still be ahead of the curve for years.

    What makes it very interesting is that SUN has re-invented itself on an Open Source CPU (SPARC) infrastructure, Open Source Firmware (OpenBoot), an Open Source Operating System (Solaris), and conforms to Open Source API's (POSIX.) Linux runs on SPARC, for those not interested to program to an open (well documented & established) API's.

    This translates into less risk than any other vendor on the market (AMD, Intel, IBM, Microsoft all require proprietary firmware, hypervisors, hardware, OS, and Applications - with the exception of Linux boot loaders or OS.)

    If SUN went out business the day after building out a data center, the source code to everything and ship it to India and have it supported indefinitely for practically no dollars.

    SUN has uniquely positioned itself to be everything to everyone (best price/performance, best performance, only completely open source stack, superb proprietary software support, and solutions from thin client, to desktop, to worktations, to server, to mainframe.)

    This is a network & systems management infrastructure support platform dream-come-true!

    The problem is... with UltraSPARC IV+ and UltraSPARC T1 - SUN may have just put themselves out of business. The market never responds well to revolutionary technology - large businesses would rather be conservative and lose money than be competitive because their pockets are so deep and they can afford to.
    • UltraSPARC IV+ and UltraSPARC T1