This is the time of year to take stock in where high-performance computing (HPC) sits and where it is headed. That's because the SC09 conference is taking place in Portland, Ore., this week and it's the biggest HPC conference around.
SC is an odd duck as conferences go. Last year it had more than 10,000 attendees and, yet, it's a largely volunteer-organized event in a world where trade shows of this scope are packaged by conference specialists or some specific corporation. Think the much-renamed LinuxWorld (run by IDG) or VMworld (run by VMware).
"SC" comes from supercomputing. Today's large computer complexes are typically not supercomputers in the sense of a specialized architecture only suitable for a specific type of technical computing. Rather, as Ashlee Vance notes in The New York Times, "The supercomputing world was long dominated by systems that required specialized chips, memory systems and networking technology. But about 10 years ago, researchers realized they could link thousands of cheaper machines running on mainstream chips and achieve pretty solid performance."
Thus an HPC event is no longer about supercomputers per se (although the term is still used as a convenient moniker for a collection of resources managed as a single entity in a single location). Rather it's about the computing components, the interconnects, the storage, and the software that ties everything together and the applications that run on top.
The Top500 nicely illustrates the evolution of HPC over time. This list, released twice annually, ranks the largest publicly acknowledged supercomputers--as the term is used today--on the basis of a somewhat simplistic, but objective, benchmark. The Top500 entries are certainly not typical of mainstream HPC; they're the biggest of the big. But they nonetheless provide some quantitative insight into important trends.
The newest iteration of the list was released Friday. There were no striking departures from the trends of the last few years, but there was some continued evolution that's worth taking note of.
The continued rise of InfiniBand. InfiniBand is a system interconnect that offers a higher performance alterative to the ubiquitous Ethernet. Although its initial backers envisioned a broader role for the technology, it's settled nicely into HPC and, to a lesser degree, back-end commercial data center functions like database clusters where low latency and high bandwidth are also paramount. (The Sun/Oracle Exadata appliance uses InfiniBand for example.)
(Credit:
TOP500.org)
InfiniBand's initial growth in HPC wasn't so much about displacing Ethernet as it was displacing the fractured collection of high-performance interconnects that preceded it. Myricom's Myrinet and Quadrics' QsNet were the most common of these, but there were many. This year InfiniBand is deployed on 181 of the Top500, a 28 percent increase from a year ago.
That's a striking increase clearly. But what is perhaps more striking is that about half that increase came at the expense of Ethernet rather than mopping up a variety of older or proprietary connection technologies. This shift started between 2007 and 2008 but was even more pronounced this year.
It's certainly possible that the next 10GbE generation of Ethernet, which today is essentially absent from the list, could again push Ethernet's numbers higher. However, whatever the specific technology, the message that I take away is that large computer clusters are starting to favor more optimized interconnects even if they and the components they connect are largely off-the-shelf.
And we see an analogous trend with the proliferation of blade servers as well. Blades, a more modular and pluggable approach to system design, have proven popular in many enterprises and midmarket companies, in part, because they help bring together computing, storage, and networking technologies into a single integrated whole. That type of integration isn't of much interest in HPC. Rather, blades play to HPC by offering high densities and reducing cable count and complexity.
In fact, among x86 servers at any rate, dominance is not too strong a word to describe the presence blades in the Top500. Consider just one vendor, Hewlett-Packard. HP has 208 ProLiant systems on the list. A full 203 of these, almost 98 percent, are ProLiant c-Class blades.
Collectively, these trends suggest what might be thought of as a trend toward building optimization around standardization. In the main, especially as one moves down from the very top of the list, the Top500 is composed mostly of systems using mainstream technologies such as x86, Linux, and standard interconnects. Clusters are the dominant architecture.
But we're increasingly not seeing mere rackmount servers connected by Gigabit Ethernet. As the systems on the Top500 list grow in capability, we're seeing more focus on how they're packaged, powered, and connected.
I plan to delve into Hewlett-Packard's new ProLiant SL Extreme Scale-Out (ExSO) line more deeply in an Illuminata Insight over the coming weeks. But it's a significant announcement that highlights some important trends, so hitting some of the highlights of today's announcement is worthwhile.
(Credit:
Hewlett-Packard)
To start off, it's a new ProLiant form factor that joins existing tower, rack, and blade lineups. Essentially, it represents a shift up to the next lot size of server purchases. In other words, tower servers came first and were often purchased one at a time. Rack servers a few at a time. Blade servers: a chassis worth, maybe 8 or 12 at a time. ProLiant SL is optimized around purchases of a rack at a time.
ProLiant SL also optimizes around the requirements of the sorts of customers who make purchases at this sort of scale. This means focusing on metrics such as performance per watt or dollar or square foot. It also means leaving out the things such customers don't care about. For example, large-scale Web and HPC sites tend to build and use their own management tools. They're looking to server vendors to mostly just provide low-level tooling for monitoring and updates.
For the ExSO debut, HP is introducing three servers that are physically a sort of horizontal blade server--though HP chose to describe them to me as "skinless servers." The servers go into a new 2U z6000 chassis that then goes into a standard rack. (Typically, five chassis at a time will go into a standard HP rack using a 10U bulk rail kit.)
- ProLiant SL160z G6 is optimized for large memory, such as applications that benefit from a large memory cache near the processing
- ProLiant 170z G6 is optimized for large storage, such as for Web search and database applications
- ProLiant SL2x170z is optimized for compute density, such as for many HPC and Web front-end applications
Although no other major vendor has quite the same design approach to high-scale x86 computing, conceptually, there is a great deal in common here with systems like IBM's iDataPlex and Sun's blades.
HP will argue that ProLiant is based on more standardized components, such as standard racks, and can more easily mix and match with third-party components. There is some truth in that assertion. However, from my perspective, what's most distinctive about this product announcement is not so much the particular hardware that HP is selling but rather its context.
Namely, this announcement extends from and builds on supply chains, channels, and the considerable success of ProLiant in the marketplace. What would be a mildly interesting server design from a smaller or less successful server vendor is very interesting coming from HP.
Few things are more annoying to PR folks than being all revved up for a big announcement, and then some big event comes along and sucks all the air out of the tech news cycle for the day.
That's what happened to Hewlett-Packard on Monday. The company announced its BladeSystem Matrix. And Oracle announced its decision to acquire Sun Microsystems.
Guess which won the coverage game. (And, yes, I'm one of the guilty parties.)
On a normal day
But on a normal day, Matrix would have, or at least should have, generated a lot of interest. Here's why.
First, a little background about server blades. They were initially pitched back around 2000 as a hardware-focused approach that disaggregated processing from the other components of a computer system and physically consolidated it. We still see remnants of that vision when blades are used in high-performance computing and some high-transaction Web sites.
But they've largely evolved in a different direction. Software (for management and virtualization) and a new style of integration are now the most important elements of mainstream commercial blade products. Many of the virtues extolled early-on for blades--fewer cables, better density, less physical redundancy--remain to be sure. But even if they're necessary, they're not sufficient.
Which brings us to HP BladeSystem Matrix. It's very explicitly an integration play. As HP put it to me: "Matrix is the box, the operating environment, the partners. The overarching environment. It's much more deeply integrated than just a bundle."
The box is HP c-Class blades (both Integrity and ProLiant) including HP's Virtual Connect 8Gb Fibre Channel and Flex-10 Ethernet modules. Virtual Connect essentially virtualizes LAN and SAN connections--allowing them to be switched among server blades inside the box in response to changing resource needs.
The software is something HP calls the Matrix orchestration environment. It brings together several components that HP already sells (and can also ship separately): Insight Dynamics-VSE, Orchestration, Recovery, and Virtual Connect Enterprise Manager (EM). Collectively, this software provides automated provisioning, capacity planning, and disaster recovery.
Finally, partner integration goes beyond the usual "we work with partners." Everyone does that. Specific application templates done in conjunction with the likes of SAP provide standardized (but customizable) ways of setting up complex software environments from a self-service portal.
Down this road before
HP has been down this road conceptually before.
In 2001, it offered up the Utility Data Center (UDC) as the ultimate adaptive, dynamic data center infrastructure--in other words, the sort of attributes it now associates with Matrix. But UDC was ahead of its time. It was tied to a lot of expensive, dedicated components. And it was not especially integrated with software outside of HP. Matrix--as well as other takes on dynamic data centers such as that from VMware--may incorporate proprietary technologies, but it's much more rooted in industry standards and standard components than UDC was. (CNET's Dan Farber has a 2004 take on UDC's demise .)
But the UDC vision and the things that HP learned from the experience clearly informed its work on Matrix.
One last note. I've seen some commentary in the vein of this announcement being a reaction to Cisco Systems. Cisco's Unified Computing System has justifiably grabbed a lot of attention. After all, Cisco is an industry heavyweight. And if Cisco is indeed serious about becoming a data center player beyond networking gear, that's big news.
However, Cisco is just starting out. HP is on its second major generation of blades and is building on software and earlier projects that, in some cases, go back a decade or more.
So to call Matrix a reaction to a newcomer gives Cisco more credit than it deserves at this point. Rather, I view this as HP continuing to build on existing plays that have already led its blades to the No. 1 spot.
This is a big week for Intel processor-based server announcements. Intel is rolling out its new "Nehalem" Xeon 5500 processor for dual-socket servers, far and away the biggest chunk of the server market by volume.
As Brooke Crothers notes on this CNET Blog Network post, "Nehalem offers some important firsts for Intel, including an integrated memory controller for better performance, hyper-threading for up to 16 virtual cores (which improves multitasking), and Turbo Boost Technology, which dynamically increases the processor's frequency (speed), as needed."
Just about any new Intel Xeon processor is paired with a spate of server announcements--after all, it's the servers that most end users buy not the chips. However, because Nehalem gives the heart of the server market such a nice performance boost, this launch is bigger than most. All the big system suppliers are making significant product introductions.
Take HP, for example. Paul Gottsegan, who leads marketing for HP's Industry Standard Servers (ISS) group, described the launch of their new ProLiant (x86) servers to me as "the biggest announcement in 20 years for ISS."
What really struck me about the HP launch though wasn't its scope--broad though it was.
Rather, it was that HP didn't take it as an opportunity to pile on an endless litany of speeds and feeds. Sure, it provided me with specifications but in the vein of supporting data rather than the core of the announcement. HP similarly focused primarily on higher-level operational and business value messages at its Technology Solutions Group (TSG) industry analyst event in Boston last week. (TSG is the business unit within HP that includes servers, software, and service.)
Unsurprisingly, in the current climate, a lot of that value message is around doing more without spending more.
The HP ProLiant G6 line's advances in energy efficiency, virtualization and automation, make it ideal for all customers. These innovations are combined with comprehensive financing programs and service offerings to redefine server economics. The new HP ProLiant G6 servers are available in 11 standards-based tower, rack and blade platforms. This represents the largest HP ProLiant rollout in company history.
"Now more than ever, customers want the best possible return on their server investments," said Christine Reischl, senior vice president and general manager, Industry Standard Servers, HP. "Building on HP's long history of hardware and software development, G6 brings together the best HP innovations in energy efficiency, virtualization and services to enable our customers to do more with less."
Now, if you haven't been a longtime HP follower as I have, the fact that technology isn't front and center may seem unremarkable. Sure, IT vendors have a proclivity to getting lost down in the weeds. But the largest and most sophisticated of those vendors--of which HP is certainly one--do understand that customers buy outcomes rather than individual products.
However, HP as a whole has perhaps struggled more than most to build on, rather than lead with, a technology message. The company is, after all, in no small part an amalgam of very engineering-centric cultures--the old HP and Digital Equipment perhaps most of all. But also Compaq and Tandem in their own ways.
This is a solid server rollout. But it's also a clear indication of HP's evolution.
I've been following the goings on at the low end of the notebook market with considerable interest over the past year. Part of my reason is professional; the way that the most mobile client devices evolve says a lot about how we will access applications and what the infrastructure running those applications will look like.
I'm also interested on a personal level. On the one hand I travel quite a bit. On the other, when I'm not traveling, I generally work out of my home office where I have a hefty desktop rig with three monitors. As a result, I value portability far more than power in my notebook given that I mostly use it for relatively lightweight Web browsing and writing while I'm on the road.
Netbooks (to use Intel's term for ultraportable PCs) have become something of a phenomenon. This hasn't been so much because they've broken new ground in notebook features. In fact, the systems that generally get lumped in that bucket today tend to skirt the edge of the full PC experience. They're explicitly intended to access Web-based applications through a browser or to run some basic productivity software locally; they're not general purpose. And they use less power-hungry, but less powerful, processors such as Intel's Atom. They're inexpensive--under $500 in most cases--especially compared to traditional road warrior notebooks that have tended to be priced at a premium relative to the notebook mainstream.
So this trend toward smaller and cheaper is interesting for a lot of reasons. What it isn't, I've argued, is a clearly distinct class of system but rather, as I argued in November:
...a temporary phenomenon that will soften over time. Memory gets denser, processors get faster, LCDs get cheaper. Some of these Moore's Law-fueled advances could indeed continue to push the entry level of the notebook market down in price... But I strongly suspect that a lot of that technical advance will also go into beefing up the capabilities of notebooks in the sort of price band that a lot of consumer electronics sell for--say, sub-$500.
We're now seeing exactly that happen.
At the Consumer Electronics Show (CES), AMD announced its Athlon Neo ("Yukon") platform, which will first appear in the HP Pavilion Dv2. Dan Ackerman on CNET describes it as:
Pitched as a kind of step-up from Netbooks, Neo provides for a little more processing muscle--at least enough to power Windows Vista. The CPU is called the Neo MV-40, runs at 1.6GHz, and comes paired with ATI Mobility Radeon HD3410 graphics.
Besides that, the Dv2 also has a fairly large keyboard for a 12-inch laptop, starts at about 3.8 pounds, and is 1.3-inches thick at its thickest point. The display is a 1,280x800 LED.
Pricing will start at $699, which puts its entry level at or a bit above the high-end of Netbooks such as the Asus Eee PC. But, at the same time, this is perhaps half the price of the more fully-featured ultraportables pitched primarily toward mobile professionals. AMD's and HP's intent here is to drive the price down on portability while still providing enough screen real estate and processing horsepower to handle things like multimedia and browser with lots of tabs open smoothly. "Good enough for the real world" is how AMD Chief Marketing Officer Nigel Dessau put it to me (with the implication that lower-end Netbooks are not.)
Intel also plans to enter this segment of the notebook market with its consumer ultra-low voltage (CULV) platform later this year.
None of this should be taken to suggest that we won't continue to also see a class of smaller, cheaper notebooks that will continue to prioritize price and mobility over a larger screen and keyboard and better performance. But it's becoming clear that such systems aren't an isolated phenomenon, but rather part of a more generalized broadening of the notebook space that includes Netbooks, 17-inch gaming machines, and everything else in between.
The idea of shipping-container-as-datacenter was floated early and loudly by Sun CEO Jonathan Schwartz. It's easy to see how he got enamored with the whole shipping container theme. For a boring steel box, the history of the shipping container is a fascinating story about labor relations, standards, and globalization. (I highly recommend Marc Levinson's book The Box: How the Shipping Container Made the World Smaller and the World Economy Bigger.)
However, it's always been a bit unclear to what degree Sun viewed its "Project Blackbox" initiative as a serious business opportunity, as opposed to just a visionary idea or an opportunity to argue that Sun uniquely gets the future of computing. A variety of other vendors have adopted a similar concept. Some of them are likewise treating it as much as a showcase as a practical solution.
However, such an attitude isn't universal--especially among those vendors who are in the business of designing datacenters as well as selling servers and other gear. For example, in June, IBM added a Portable Modular Data Center (PMDC) offering within the context of a variety of modular form factors for different customer types and requirements.
The latest announcement comes from HP, which has now introduced a Performance Optimized Data Center (POD). It, perhaps most of all, emphasizes the practical and evolutionary from a technology perspective, and alignment with the needs of specific customer segments in the marketing.
Technology first. Most of the other such products on the market introduce various clever schemes to cram lots of computer, storage, networking, power, and cooling gear into the tight confines of a shipping container without compromising the ability to service and reconfigure that equipment. Not easy, and definitely not in keeping with the way access is handled in a regular bricks and mortar datacenter.
HP takes a different approach. It uses 19"-wide, full-depth racks in a 50U height. In other words, standard width, standard depth, and just a bit higher than the 42U racks that are standard in most datacenters. This should make it straightforward to install most standard IT gear--whether from HP or someone else.
The catch is that accessing the rear of the racks requires opening doors built into the side of the shipping container. That's another door to be properly sealed and a more involved procedure to get at the back of the racks. Nonetheless, it seems a reasonable tradeoff for a lot of uses--especially as it helps contribute to a very high density design as well.
From a marketing and positioning perspective, HP is positioning this product most of all for capacity expansion of existing facilities when more computing horsepower is needed in a hurry; it says that one of these PODs can be delivered in about six weeks. Closely related is what HP describes as "transitional" use--that is, capacity needed for certain window of time; appropriate financial terms and conditions, such as leases, are important in this case. Leasing also plays into "generational" deployments where customers may want to swap out a computer infrastructure en masse for a technology refresh.
Interestingly, a lot of the "sexy" container scenarios are secondary or missing in HP's plans. Extreme-density scale-out is a POD target, but one that HP sees as a smaller opportunity than use cases like capacity expansion or disaster recovery. And HP isn't especially going after military and telecommunications use cases. It says that density isn't a big deal in these segments that, instead, emphasize customization and robustness. In other words, not a good match with an HP strategy that is fundamentally about leveraging its strengths in volume server design and supply chains.
And that's the reason HP's likely to be as successful with this type of product as anyone--if not more than most. It has the datacenter design expertise, sure. EYP Mission Critical Facilities, which HP bought in 2007, brought with it truly premiere capabilities in this regard. However, it's the IT gear within the container, how it's delivered, how it's serviced, and how it's upgraded that matter most to potential customers. Those are all things HP does well. And even if HP puts great stock in its ability to integrate just about any third-party equipment, it clearly hopes that its own systems will be a major part of the mix. The rest is just a box. Even if the box did "make the world smaller and the world economy bigger."
The AdvFS file system has had a bittersweet history. When introduced by Digital Equipment Corporation in the mid-1990s for DEC's Unix flavor (which would eventually become Tru64 UNIX), AdvFS was one of the most capable--if not the most capable--Unix file systems in existence.
As a 1999 Illuminata research note by my colleague Jonathan Eunice said: "Tru64 UNIX has the features one expects of a top-drawer enterprise Unix: multiprocessing, multithreading, disk volume management, a journaling file system, multi-path I/O, extensive TCP/IP networking, and conformance with a range of de jure and de facto standards. Its AdvFS filesystem, 64-bit and large memory support, and Memory Channel cluster interconnect provide a tour de force in high-scale system design." (See also this 1996 Unix scorecard.)
However, Tru64--along with the Alpha processor on which it ran--was one of the casualties when HP purchased Compaq (which had earlier absorbed DEC). AdvFS itself seemingly got a reprieve when HP decided to port it to HP-UX along with its associated, and equally well-regarded, TruCluster clustering environment. However, in 2004, HP killed the delayed project and went with the third-party Veritas Storage Foundation instead.
Today, AdvFS gets yet another life as HP makes the source code for AdvFS available to the Open Source community. Specifically, according to the press release, "HP will contribute the code as a reference implementation of an enterprise Linux file system under the terms of General Public License Version 2 for compatibility with the Linux kernel, as well as provide design documentation, test suites and engineering resources."
That's a fair bit of history I know. However, I thought at least a little context was necessary before diving under the covers.
What's being contributed? As HP says, the source code for AdvFS and related resources. This includes both the original "mature" version for Tru64 and the version for the worked-on port to HP-UX, which included some improvements and updates. HP describes the latter as a "feature complete port." However, I take that as meaning that 90 percent of the development work is done, so that only "the other 90 percent" of development, test, QA, tuning, and so forth is left to do. (After all, if it were almost ready to go into production, presumably HP would have completed the project rather than gone the Veritas route.)
What's not being contributed? This announcement strictly concerns the Tru64 file system; it does not include TruCluster--which builds atop Tru64, but is a separate product.
When will AdvFS be available as a file system for Linux? Short answer: probably never. One Wikipedia contributor amusingly opines: "anyone should be able to port it to the Linux kernel and get it into mainline." However, the reality is that although AdvFS code may assist with and its design documents may inform future Open Source file system work, it's highly unlikely that AdvFS qua AdvFS will be plopped into Linux in its current form.
What file system project(s) could make use of AdvFS? In the relatively near-term, ext4 is the next new file system that we're likely to see widely deployed on Linux. It's a largely incremental enhancement to the ubiquitous ext3 that focuses on larger file and file system sizes as well as various performance tweaks. Versions of ext4 are starting to appear in community releases such as Fedora 9. HP, among many others, has been involved in ext4 development, but AdvFS itself won't have a big impact here. Rather, it's Btrfs (pronounced "butter f s") with which HP is looking to hookup AdvFS.
What is Btrfs? In general terms, think of Btrfs as a "next generation file system," which is to say one that moves beyond the administration and availability models used by current file systems. Although the specifics are quite different, Sun's ZFS is another example. From its project page on kernel.org: "Btrfs is a new copy on write filesystem for Linux aimed at implementing advanced features while focusing on fault tolerance, repair and easy administration. Initially developed by Oracle, Btrfs is licensed under the GPL and open for contribution from anyone."
What's HP's interest? HP has contributed to a wide range of Open Source projects over time, but this goes beyond generic community goodness. Oracle kicked off Btrfs but is apparently interested in moving it beyond being just "an Oracle filesystem project." As for HP, AdvFS (and TruCluster) were long part-and-parcel of a close development relationship with Oracle involving database clustering. It's pretty clear to me that Oracle wanted more companies involved with Btrfs development; AdvFS was a highly relevant technological bauble for HP to bring as a housewarming present for one of its most important partners.
Bottom line: Especially given the great success that its ProLiant and BladeSystem lines have been enjoying, HP has a vested interest in the continued success of Linux and its ability to scale to larger and larger workloads. A lot of that growth is about rapidly growing data, so a next generation file system is going to be needed at some point. Btrfs per se is speculative, with production deployments even in the 2011 or 2012 timeframe seeming optimistic given the slow cadence of file system roll outs. (It's customers' data we're talking about here, after all. Sun took a good couple of years to start seriously rolling out ZFS even after it was nominally "ready.") But making AdvFS resources available gives Open Source developers the opportunity for useful insights into how a sophisticated production file system should work. Especially to the degree that AdvFS wins some points with a major partner, it's a nice way to leverage some IP that would otherwise mostly go to waste.
One of the curses, I suppose, of knowing one's high-tech history is that way too many news items cause me to go, "Here we go again!"
The proximate tidbit this time is, of course, the news that Hewlett-Packard is acquiring services giant EDS for $13.9 billion. Various news organizations had previously pegged the deal value between $12 billion and $13 billion. The New York Times described it at a $12.6 billion cash transaction.
When we last saw this play, it was with Carly Fiorina in the role of HP's chief executive, looking to spend a reported $17 billion to $18 billion on PricewaterhouseCoopers Consulting in 2000. A lousy set of quarterly results turned in by HP helped to scotch that deal. It also didn't help that a lot of observers thought that HP was offering way too much for an organization with $6.7 billion in annual revenues (2001) and about 33,000 employees.
IBM seemingly provided evidence of this view when it bought PwCC in 2002 for only about $3.5 billion. (A bit of an unfair comparison, given the economic and other events of 2001, but still.) Carly went on to get her acquisition kicks by gobbling up Compaq instead.
So what, if anything, is different this time around?
The money. I'll leave the detailed financial modeling to the appropriate specialists, but here are some back-of-the-envelope numbers. In 2000, HP was looking to pay more than two times the annual revenues for PwCC, which IBM ended up getting for about 0.6 times revenues instead. In this case, HP spent less money ($13.9 billion) for a larger ($22.1 billion annual revenues) organization. At least by this measure, HP's expenditure is therefore much more in line with what IBM eventually spent for PwCC than it is what HP had initially proposed.
HP management capabilities. Especially after this acquisition, something that's really striking is just how closely HP has maintained the course that Carly laid out. There's a slight difference, of course. If one goes back a few years, the boat may have been on a sensible bearing, but it was springing leaks in just about every compartment.
Carly has argued that post-Compaq financial problems just needed more time to work themselves out. Perhaps--but I'm skeptical. In any case, Mark Hurd has made remarkably few changes to HP's strategic direction since he took over. The benefits of scale promised from the Compaq buy have indeed proven out. EDS represents growth of scale along another axis--services--that puts HP that much more in the mold of IBM. The difference from times past is that Mark has a track record for keeping things ship-shape.
HP has made services acquisitions before, but they've been targeted and specialized. The most recent was of EYP Mission Critical Facilities, a data center design specialist that gives HP some legitimate differentiation in the power and cooling game. EDS is broader and bigger than even PwCC would have been.
On the one hand, this means a lot of personnel and fixed costs of the sort that have been no small issue for IBM--the company HP is attempting to mirror in important ways. On the other hand, if you believe--as I do--that companies (especially in small and midsize businesses) are increasingly going to move their computing off their premises and into data centers run by specialists, then acquiring the sort of large-scale hosted services business that EDS includes among its many operations isn't a bad direction for a system supplier at all.
With the ProLiant DL785 G5 Server, Hewlett-Packard has re-entered the 8-socket x86 server space. This system has twice the computing headroom of the quad-processor servers that are generally considered at the top end of the volume or so-called commodity server space.
HP isn't new to this market segment. In 1997, Intel bought a company by the name of Corollary that was in the process of developing a chipset that effectively "glued together" two standard quad-processor x86 busses into a single 8-way symmetrical multiprocessor (SMP). Intel not only completed development, it also gave the chipset legitimacy by giving it an Intel blaze. Then Microsoft provided the last major missing piece with Windows 2000, an OS that not only showed real progress in reliability and scalability over its predecessors, but also lent credibility to Microsoft's efforts to be perceived as a serious OS vendor for serious servers.
ProLiant, initially as a Compaq server brand and then after its acquisition by HP, used this chipset and its successors for a succession of server products--even after Intel decided to stop contributing to further development. (Intel had, at various points, planned to do a Xeon version of Itanium's 870 chipset, but this never ended up happening.) Compaq's own version, the "ProLiant F8" chipset, adapted Profusion for the architecture and bus speeds associated with newer Intel processors, but did not fundamentally alter the design. (Subscribers can read about more of the historical background here.)
However, HP eventually decided to pull the plug on in-house development of 8-way chipsets for Xeon. I've broached the question "Why?" with HP executives on a number of occasions over the past few years and their responses have been pretty consistent. They've boiled down to two basic rationales:
- The size of the 8-socket market does not justify the expense associated with custom chipset development.
- To the degree that there's a demand for larger ProLiants, it's mostly from customers wanting to run larger Microsoft SQL Server databases and associated enterprise applications--and those needs can be met by Windows running on HP Integrity (Itanium-based) servers.
So what's changed to bring ProLiant back into this space? From my perspective, there's probably not one single reason but rather a few different factors that collectively served the needle from "No" to "Yes."
It's easier. Rather than using Intel processors, the ProLiant DL785 G5 uses Advanced Micro Devices Opteron "Barcelona" quad-core processors. Unlike Xeons, the AMD processors can support up to 8-socket servers without the use of special server vendor-developed chips. A lot of effort (and therefore money) still goes into designing, qualifying, and supporting a system in this class. However, the costs associated with primarily integrating existing in-house and third-party components and technologies are still much less than if bespoke chipset design is added to the mix.
The market is larger. Dual-socket servers still make up the bulk of server unit sales. However, server virtualization, in particular, has kicked demand for larger boxes, which once seemed to be on an inevitable slide, up a notch. Server virtualization allows as many workloads (more or less) to run on a system as processor, memory, and I/O capacity can support. Given this, many users are starting to think that they're better off consolidating onto larger servers than smaller ones. This reduces the number of physical boxes to manage. In addition, larger servers often come with a more sophisticated array of reliability and management features. The market for scale-up x86 servers isn't going away either--for reasons including the increasing sophistication of Microsoft SQL Server or the growth of Solaris on x86.
Integrity is only a partial solution. From HP's perspective, the "buy Itanium" message was always logical enough. Most of the critical high-end Windows applications were available and Integrity, after all, was specifically optimized for that space. It's a good story, but the reality is that a lot of Windows customers don't want to support multiple processor architectures in their environments--even if the software is (mostly) the same.
As a final point, the HP of today is a tightly managed and highly measured organization. And ProLiant is clearly one of the growth stars. Thus, it's not hard to imagine that politely leaving high-end Windows opportunities to Integrity came to be regarded as sub-optimal from the perspective of HP as a whole.
Whatever the precise balance of reasons, HP is back in the 8-socket Xeon game. It's a space that HP has largely ceded to IBM's X4 designs. Now HP is re-engaging aggressively as they did with blades and as they've done across so much of the x86 space.
Sometime around 1990, Data General (who I worked for at the time) came out with a portable terminal called the Walkabout. The idea was that it would let people check their e-mail from the road using the built-in modem and terminal emulator, while being lighter and cheaper than the portable computers of the day. It wasn't as silly an idea as it might seem today--lots of people still used terminals rather than PCs at the time--but, like the DG/One, it was ahead of the hardware curve, and pricey.
Fast forward to 2007. Palm announced the Foleo as a companion to the Treo smartphone. Palm founder Jeff Hawkins called this portable "Internet interface appliance" the best idea he ever had. Many others, including myself, begged to differ and, in the end, the product never made it to market. The issues hadn't changed all that much in close to 20 years: too similar in price and bulk to a full laptop, and a problematic fit for an only somewhat connected world.
Yet here's Hewlett-Packard introducing its first mobile thin client, a fruit of its 2007 acquisition of Neoware. From the press release:
Designed for mobile workers who have a well-defined set of tasks requiring a wide range of general-purpose software or unique business applications, the HP Compaq 6720t Mobile Thin Client is ideal for on-the-go professionals such as insurance claim processors, remote staff, warehouse and inventory managers and office administrators.
It is based on Microsoft Windows XPe and features a 15.4-inch display; solid-state design with no hard drive, fan or other moving parts; enhanced security with no data residing on the notebook; Wi-Fi Certified WLAN along with support for 3G broadband wireless via PC memory card slot; and solid-state flash module for greater durability, faster data access and more quiet and cool operation.
The 6720t also helps increase security by accessing software applications hosted on a server, virtual PC or blade PC computing platform over a secure virtual private network Internet connection. Data files and software applications also are saved remotely on a secure server to help reduce the risk of data loss, viruses and product theft.
Client management is simplified, as IT administrators are able to remotely install, manage, update and execute application software simultaneously across an entire fleet of clients that are pushed to the mobile thin clients as soon as they are connected to the network.
So why does this make sense when the earlier examples I cited didn't? There are a few factors to consider.
The general interest in thin clients. Businesses are showing an interest in thin clients as a category that far exceeds what we've seen historically. There are a variety of reasons for this that I cover in Desktops on Diets, but essentially we're seeing a convergence of trends that are achieving a certain critical mass. As a result, even though mobile thin clients may be just a slice of the overall thin client market, that overall pie is growing rapidly.
Connectivity isn't ubiquitous, but it is reliably available in certain environments. Wired campuses are becoming commonplace--universal even. More and more people likewise have wireless connectivity throughout their homes. In other words, workers who are only using their notebooks at work and at home actually do have essentially ubiquitous broadband connectivity. HP estimates that 35 percent to 40 percent of notebook users closely match this description--that is, they may take their notebooks with them to meetings or to do some work in the evening but they don't typically take them on the road.
And the management and security advantages of thin clients can be considerable. Especially when they're mobile devices. Indeed, although it's true that mixing "thin" (as in stateless) and "mobile" serves up its own set of challenges, the potential benefits can actually be greater than with desktops. After all, a desktop PC isn't likely to get left in the back of a taxicab somewhere and the fact that desktop PCs are always tethered to an Ethernet cable simplifies pushing out application updates.
HP isn't the first vendor to offer a mobile thin client. But the other entrants are mostly small firms. This announcement is notable because it isn't just another point device. Rather, it's a mobile thin client that's not only part of a broad lineup of thin client devices, but is also part of a broad lineup of HP hardware, software, and services offerings such as Consolidated Client Infrastructure (CCI). And that's really an important storyline in the rise of thin clients.
One size does not fit all, but historically vendors tended to push the narrow approaches in which they specialized. That's changing today as companies like Citrix, HP, and IBM are combining technologies developed in-house and through partners to assemble a broader application delivery story that isn't about using the same hammer for every problem.
