Category Archives: Open Compute Project

Inevitability of Virtualized Infrastructure

As a previous post, Infrastructure Virtualization Versus Converged Infrastructure, attests, I strongly believe that virtualization is leading us to a future in which underlying hardware becomes largely undifferentiated and interchangeable. Applications and orchestration will reside in software riding atop the virtualization layer, which effectively will function as an abstraction buffer above hardware infrastructure.  The latter will eventually include hardware for computer, networking, and storage.

Vendors that ride hardware-based business models will have trouble adapting to this new reality. Many of these companies have hordes of software developers and software engineers, but they inextricably intertwine their software and hardware as a matter of business practice, selling the latter as proprietary boxes that often cannot interoperate with, or be swapped out for, competing hardware. It’s classic hardware-based vendor lock-in, and it’s been with us for many years. This applies to vendors that sell all three main types of hardware infrastructure, and to those that sell them tied together as converged infrastructure.

Loosening a Tenacious Grip

Proprietary data-center hardware would appear to be running on borrowed time, though it will not disappear overnight. Its grip will be especially tenacious in the enterprise, though the pull of the cloud eventually will weaken its hold. Proprietary compute infrastructure will be the first to succumb, but networking and storage will fall, too. The economic and operational logic powering the transition is inexorable, so it’s a question of when, not whether, it will happen.

While CapEx cost savings are an obvious benefit, operational flexibility (shifting workloads with agility and less effort) and OpEx savings also are factors. Infrastructure hardware will be cheaper, as well as easier and less costly to run. Pools of industry-standard hardware will be reallocated on demand to serve the needs of application workloads. Data-center customers no longer will be constrained by the hardware-release schedules of their previous vendors of choice. Customers also will be able to take advantage of the latest industry-standard chipsets, which will power hardware with improved energy efficiency and better cooling characteristics.

In servers, and now in storage, Facebook’s Open Compute Project (OCP) has sought to expedite the move to off-the-shelf hardware. Last week at Oscon, Frank Frankovsky, a vice president at  Facebook and the chairman and president of the OCP, rallied the open-source troops by arguing that proprietary x86 systems are “gratuitously differentiated.” He called for all hardware-design specifications to be open.

OCP as Competitive Cudgel

That would benefit Facebook, which launched OCP as a vehicle to help it lower data-center CapEx and OpEx, boost operational flexibility, and — last but not least — mitigate a competitive advantage held by Google, which had a massive head start in rationalizing and fine-tuning its data centers and IT infrastructure. In fact, Google cloaks its IT operations in extreme secrecy, believing that its practices and technologies deliver substantial competitive advantage over its main rivals, including Facebook. The latter must agree, because the animating idea behind Open Compute is to create a market, demand and supply, for commodity server hardware will reduce or eliminate Google’s edge.

Some have wondered why Google hasn’t joined OCP, but the answer should be obvious. Google believes it has cracked the infrastructure code, and it is therefore disinclined to share its insights and best practices with its competitors. Google isn’t a fan of proprietary vanity hardware — it’s been designing its own gear, then going to server and network ODMs, for some time now — but Google feels it has nothing to gain, and much to lose, from opening its kimono to the OCP crowd.

With networking, though, Google felt it needed a little help from its friends — as well as from its enemies. That explains why it allied with Facebook and other cloud-service providers in the Open Networking Foundation (ONF), which I have written about here on many occasions. The goal of the ONF, as with OCP, is to slip the proprietary shackles of hardware vendors, whose gear functions as an impediment to operational agility as well as a costs that could be reduced through SDN-style network virtualization. Google’s communitarian approach to addressing the network-virtualization riddle suggests that it believes it cannot achieve the desired outcome on its own.

Cracking the Nut

Whereas compute hardware was well on its way to standardization, networking hardware, until the ONF, was akin to a vertically integrated mainframe system, replete with a proliferating number of both proprietary and industry-standard protocols. Networking is a bigger, and tougher, nut to crack.

But crack it will, first at the big cloud-service providers, then, as the cloud gains momentum, at enterprises.

PS: I will post something tomorrow about VMware’s just-announced acquisition of Nicira, which is big news no matter how you slice it.  I wrote the above post before I learned of the acquisition.

Advertisements

Like OpenFlow, Open Compute Signals Shift in Industry Power

I’ve written quite a bit recently about OpenFlow and the Open Networking Foundation (ONF). For a change of pace, I will focus today on the Open Compute Project.

In many ways, even though OpenFlow deals with networking infrastructure and Open Compute deals with computing infrastructure, they are analogous movements, springing from the same fundamental set of industry dynamics.

Open Compute was introduced formally to the world in April. Its ostensible goal was “to develop servers and data centers following the model traditionally associated with open-source software projects.”  That’s true insofar as it goes, but it’s only part of the story. The stated goal actually is a means to an end, which is to devise an operational template that allows cloud behemoths such as Facebook to save lots of money on computing infrastructure. It’s all about commoditizing and optimizing the operational efficiency of the hardware encompassed within many of the largest cloud data centers that don’t belong to Google.

Speaking of Google, it is not involved with Open Compute. That’s primarily because Google has been taking a DIY approach to its data center long before Facebook began working on the blueprint for the Open Compute Project.

Google as DIY Trailblazer

For Google, its ability to develop and deliver its own data-center technologies — spanning computing, networking and storage infrastructure — became a source of competitive advantage. By using off-the-shelf hardware components, Google was able to provide itself with cost- and energy-efficient data-center infrastructure that did exactly what it needed to do — and no more. Moreover, Google no longer had to pay a premium to technology vendors that offered products that weren’t ideally suited to its requirements and that offered extraneous “higher-value” (pricier) features and functionality.

Observing how Google had used its scale and its ample resources to fashion its cost-saving infrastructure, Facebook  considered how it might follow suit. The goal at Facebook was to save money, of course, but also to mitigate or perhaps eliminate the infrastructure-based competitive advantage Google had developed. Facebook realized that it could never compete with Google at scale in the infrastructure cost-saving game, so it sought to enlist others in the cause.

And so the Open Computer project was born. The aim is to have a community of shared interest deliver cost-saving open-hardware innovations that can help Facebook scale its infrastructure at an operational efficiency approximating Google’s. If others besides Facebook benefit, so be it. That’s not a concern.

Collateral Damage

As Facebook seeks to boost its advertising revenue, it is effectively competing with Google. The search giant still derives nearly 97 percent of its revenue from advertising, and its Google+ is intended to distract it not derail Facebook’s core business, just as Google Apps is meant to keep Microsoft focused on protecting one of its crown jewels rather than on allocating more corporate resources to search and search advertising.

There’s nothing particularly striking about that. Cloud service providers are expected to compete against other by developing new revenue-generating services and by achieving new cost-saving operational efficiencies.  In that context, the Open Compute Project can be seen, at least in one respect, as Facebook’s open-source bid to level the infrastructure playing field and undercut, as previously noted, what has been a Google competitive advantage.

But there’s another dynamic at play. As the leading cloud providers with their vast data centers increasingly seek to develop their own hardware infrastructure — or to create an open-source model that facilitates its delivery — we will witness some significant collateral damage. Those taking the hit, as is becoming apparent, will be the hardware systems vendors, including HP, IBM, Oracle (Sun), Dell, and even Cisco. That’s only on the computing side of the house, of course. In networking, as software-defined networking (SDN) and OpenFlow find ready embrace among the large cloud shops, Cisco and others will be subject to the loss of revenue and profit margin, though how much and how soon remain to be seen.

Who’s Steering the OCP Ship?

So, who, aside from Facebook, will set the strategic agenda of Open Compute? To answer to that question, we need only consult the identities of those named to the Open Compute Project Foundation’s board of directors:

  • Chairman/President – Frank Frankovsky, Director, Technical Operations at Facebook
  • Jason Waxman, General Manager, High Density Computing, Data Center Group, Intel
  • Mark Roenigk, Chief Operating Officer, Rackspace Hosting
  • Andy Bechtolshiem, Industry Guru
  • Don Duet, Managing Director, Goldman-Sachs

It’s no shocker that Facebook retains the chairman’s role. Facebook didn’t launch this initiative to have somebody else steer the ship.

Similarly, it’s not a surprise that Intel is involved. Intel benefits regardless of whether cloud shops build their own systems, buy them from HP or Dell, or even get them from a Taiwanese or Chinese ODM.

As for the Rackspace representation, that makes sense, too. Rackspace already has OpenStack, open-source software for private and public clouds, and the Open Compute approach provides a logical hardware complement to that effort.

After that, though, the board membership of the Open Compute Project Foundation gets rather interesting.

Examining Bechtolsheim’s Involvement

First, there’s the intriguing presence of Andy Bechtolsheim. Those who follow the networking industry will know that Andy Bechtolsheim is more than an “industry guru,” whatever that means. Among his many roles, Bechtolsheim serves as the chief development officer and co-founder of Arista Networks, a growing rival to Cisco in low-latency data-center switching, especially at cloud-scale web shops and financial-services companies. It bears repeating that Open Compute’s mandate does not extend to network infrastructure, which is the preserve of the analogous OpenFlow.

Bechtolsheim’s history is replete with successes, as a technologist and as an investor. He was one of the earliest investors in Google, which makes his involvement in Open Compute deliciously ironic.

More recently, he disclosed a seed-stage investment in Nebula, which, as Derrick Harris at GigaOM wrote this summer, has “developed a hardware appliance pre-loaded with customized OpenStack software and Arista networking tools, designed to manage racks of commodity servers as a private cloud.” The reference architectures for the commodity servers comprise Dell’s PowerEdge C Micro Servers and servers that adhere to Open Compute specifications.

We know, then, why Bechtolsheim is on the board. He’s a high-profile presence that I’m sure Open Compute was only too happy to welcome with open arms (pardon the pun), and he also has business interests that would benefit from a furtherance of Open Compute’s agenda. Not to put too fine a point on it, but there’s an Arista and a Nebula dimension to Bechtolsheim’s board role at the Open Compute Project Foundation.

OpenStack Angle for Rackspace, Dell

Interestingly, the presence of Bechtolsheim and Rackspace’s Mark Roenigk on the board both emphasize OpenStack considerations, as does Dell’s involvement with Open Compute. Dell doesn’t have a board seat — at least not according to the Open Compute website — but it seems to think it can build a business for solutions based on Open Compute and OpenStack among second-tier purveyors of public-cloud services and among those pursuing large private or hybrid clouds. Both will become key strategic markets for Dell as its SMB installed base migrates applications and spending to the cloud.

Dell notably lost a chunk of server business when Facebook chose to go the DIY route, in conjunction with Taiwanese ODM Quanta Computer, for servers in its data center in Pineville, Oregon. Through its involvement in Open Compute, Dell might be trying to regain lost ground at Facebook, but I suspect that ship has sailed. Instead, Dell probably is attempting to ensure that it prevents or mitigates potential market erosion among smaller service providers and enterprise customers.

What Goldman Sachs Wants

The other intriguing presence on the Open Compute Project Foundation board is Don Duet from Goldman Sachs. Here’s what Duet had to say about his firm’s involvement with Open Compute:

“We build a lot of our own technology, but we are not at the hyperscale of Google or Facebook. We are a mid-scale company with a large global footprint. The work done by the OCP has the potential to lower the TCO [total cost of ownership] and we are extremely interested in that.”

Indeed, that perspective probably worries major server vendors more than anything else about Open Compute. Once Goldman Sachs goes this route, other financial-services firms will be inclined to follow, and nobody knows where the market attrition will end, presuming it ends at all.

Like Facebook, Goldman Sachs saw what Google was doing with its home-brewed, scale-out data-center infrastructure, and wondered how it might achieve similar business benefits. That has to be disconcerting news for major server vendors.

Welcome to the Future

The big takeaway for me, as I absorb these developments, is how the power axis of the industry is shifting. The big systems vendors used to set the agenda, promoting and pushing their products and influencing the influencers so that enterprise buyers kept their growth rates on the uptick. Now, though, a combination of factors — widespread data-center virtualization, the rise of cloud computing, a persistent and protected global economic downturn (which has placed unprecedented emphasis on IT cost containment) — is reshaping the IT universe.

Welcome to the future. Some might like it more than others, but there’s no going back.

HP’s Launches Its Moonshot Amid Changing Industry Dynamics

As I read about HP’s new Project Moonshot, which was covered extensively by the trade press, I wondered about the vendor’s strategic end game. Where was it going with this technology initiative, and does it have a realistic likelihood of meeting its objectives?

Those questions led me to consider how drastically the complexion of the IT industry has changed as cloud computing takes hold. Everything is in flux, advancing toward an ultimate galactic configuration that, in many respects, will be far different from what we’ve known previously.

What’s the Destination?

It seems to me that Project Moonshot, with its emphasis on a power-sipping and space-saving server architecture for web-scale processing, represents an effort by HP to re-establish a reputation for innovation and thought leadership in a burgeoning new market. But what, exactly, is the market HP has in mind?

Contrary to some of what I’ve seen written on the subject, HP doesn’t really have a serious chance of using this technology to wrest meaningful patronage from the behemoths of the cloud service-provider world. Google won’t be queuing up for these ARM-based, Calxeda-designed, HP-branded “micro servers.” Nor will Facebook or Microsoft. Amazon or Yahoo probably won’t be in the market for them, either.

The biggest of the big cloud providers are heading in a different direction, as evidenced by their aggressive patronage of open-source hardware initiatives that, when one really thinks about it, are designed to reduce their dependence on traditional vendors of server, storage, and networking hardware. They’re breaking that dependence — in some ways, they see it as taking back their data centers — for a variety of reasons, but their behavior is invariably motivated by their desire to significantly reduce operating expenditures on data-center infrastructure while freeing themselves to innovate on the fly.

When Customers Become Competitors

We’ve reached an inflection point where the largest cloud players — the Googles, the Facebooks, the Amazons, some of the major carriers who have given thought to such matters — have figured out that they can build their own hardware infrastructure, or order it off the shelf from ODMs, and get it to do everything they need it to do (they have relatively few revenue-generating applications to consider) at a lower operating cost than if they kept buying relatively feature-laden, more-expensive gear from hardware vendors.

As one might imagine, this represents a major business concern for the likes of HP, as well as for Cisco and others who’ve built a considerable business selling hardware at sustainable margins to customers in those markets. An added concern is that enterprise customers, starting with many SMBs, have begun transitioning their application workloads to cloud-service providers. The vendor problem, then, is not only that the cloud market is growing, but also that segments of the enterprise market are at risk.

Attempt to Reset Technology Agenda

The vendors recognize the problem, and they’re doing what they can to adapt to changing circumstances. If the biggest web-scale cloud providers are moving away from reliance on them, then hardware vendors must find buyers elsewhere. Scores of cloud service providers are not as big, or as specialized, or as resourceful as Google, Facebook, or Microsoft. Those companies might be considering the paths their bigger brethren have forged, with initiatives such as the Open Compute Project and OpenFlow (for computing and networking infrastructure, respectively), but perhaps they’re not entirely sold on those models or don’t think they’re quite right  for their requirements just yet.

This represents an opportunity for vendors such as HP to reset the technology agenda, at least for these sorts of customers. Hence, Project Moonshot, which, while clearly ambitious, remains a work in progress consisting of the Redstone Server Development Platform, an HP Discovery Lab (the first one is in Houston), and HP Pathfinder, a program designed to create open standards and third-party technology support for the overall effort.

I’m not sure I understand who will buy the initial batch of HP’s “extreme low-power servers” based on Calxeda’s EnergyCore ARM server-on-a-chip processors. As I said before, and as an article at Ars Technica explains, those buyers are unlikely to be the masters of the cloud universe, for both technological and business reasons. For now, buyers might not even come from the constituency of smaller cloud providers

Friends Become Foes, Foes Become Friends (Sort Of)

But HP is positioning itself for that market and to be involved in those buying decisions relating to the energy-efficient system architectures.  Its Project Moonshot also will embrace energy-efficient microprocessors from Intel and AMD.

Incidentally, what’s most interesting here is not that HP adopted an ARM-based chip architecture before opting for an Intel server chipset — though that does warrant notice — but that Project Moonshot has been devised not so much as to compete against other server vendors as it is meant to provide a rejoinder to an open-computing model advanced by Facebook and others.

Just a short time ago, industry dynamics were relatively easy to discern. Hardware and system vendors competed against one another for the patronage of service providers and enterprises. Now, as cloud computing grows and its business model gains ascendance, hardware vendors also find themselves competing against a new threat represented by mammoth cloud service providers and their cost-saving DIY ethos.