Mike Pav

Austin, Texas-based startup Spanning has embraced the concept of cloud computing so much that its product is a backup service for Google Apps — completely hosted and run from Amazon Web Services. The idea of backing up one cloud service via another was intriguing enough that I asked Mike Pav, the VP of engineering at Spanning, how he does it.

Spanning charges people or businesses a year to back up Google Apps, including email and documents, against the user somehow deleting or losing them. Google will support users if it loses their data, but it won’t go searching for your files if you mess up.

Spanning CEO Charlie Wood is confident enough that Google won’t get into extended services like backup that he’s sticking with this, although he’s also looking for new lines of business as the company continues to grow. To that end, the company is seeking its next round of funding after having raised a million Series A round last April.

Building a backup cloud in the cloud

Building a cloud-based backup for a cloud service requires a devotion to reliability and planning for worst-case scenarios. Creating a backup service in Amazon Web services is never done, said Pav, as he explained some of the techniques he’s used to support Spanning while also trying to keep costs in line. “For example, a single point of failure for us was our database, but we just finished up a big project to partition our database,” Pav said. “We have to focus on the path and not the destination, because as far as scalability is concerned, we’ll never be done. That’s our real barrier to entry.”

The Spanning engineering team.

Spanning adds terabytes of storage each month, and it uses Amazon because it makes automatic scaling seamless. “It would be terrible if we had to rack our own drives into an array to deal with that,” Pav says. Spanning stores all the content on S3 because it guarantees high reliability, but the getting the data to S3 can be slow. To address this, Spanning uses parallel access, which helps address the speed of S3, but also provides an added benefit in terms of scalability and reliability.

Designing messaging so dying VMs won’t take out your data

Spanning uses Amazon SQS to queue work to a pool of virtual resources that grows and shrinks based on load. Pav’s team has set up Spanning’s application to track the incoming flow of data to EC2 and make sure each time the system is about to back up new content, it checks to see if the EC2 instance is about to shut down. If it is, the in-progress backup requeues its work-in-progress so another server can pick up this work when AWS adds another server from the pool. That way, the backup doesn’t have to start all over again.

This is important when dealing with potentially large sets of data. Pav says Amazon offers several different models for queue management, but simplicity and scalability are the driving features for Spanning. “When you’re dealing with large data sets for a large number of users, you can’t afford to do anything twice.”

Don’t do anything Amazon will do for you

From storing papers to storing packets.

Spanning uses Amazon Relational Database (RDS) for its persistent database storage, although it does impose limitations on how much data Spanning can store and the throughput it can support on any single database instance. Pav admits this limits his partitioning strategies, but he’s willing to work within those limits, because it cuts his need to support and build his own data store.

“We want to get out of the business of spending time managing these things. We can solve this problem at the application’s architectural level to make sure it scales,” he said. “RDS may not be the highest-performance option, but we are able to reduce investment into something that’s not core to our business and by making good application level architectural decisions we can render the RDS performance issue moot.”

Amazon has changed not just the economics of building an IT service, but also helps make his product better and faster at less cost to him and his team. Pav notes that because of the reliability of Spanning on Amazon and his confidence that user data won’t be lost, he deploys new code when features are ready, and often in the middle of the day when his team is fresh. This is a big shift from the older days of waiting until late at night when theoretically fewer users are online to feel any disruptions.

Of course, with a large customer base all over the world and a growing one in North America, Pav points out that in today’s distributed world, there really is no more middle of the night.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• A field guide to cloud computing: current trends, future opportunities
• Infrastructure Q3: OpenStack and flash step into the spotlight
• What Amazon’s new Kindle line means for Apple, Netflix and online media

alt=''
border='0'
/>

GigaOM

Servers? We don't need no stinkin' servers!

Both mobile and high-performance computing are placing huge power efficiency and performance demands on chips, but the ,000-question is how long until such extreme computing use cases hit the server mainstream. Asked another way, the question becomes, how long until AmazonWeb Services adopts ARM-basedservers?

Or perhaps it isn’t ARM-based servers, but a variation on an Intelchip that takes its architecture from some of the more innovative and energy-efficient silicon out today. For example, Adapteva, a startup I profiled back in May, on Monday released a 64-core chip that can deliver 70 gigaflops of performance per watt. If you don’t speak gigaflops, that’s okay: It basically has done what Intel and certain countries have deemed impossible with the current generation of silicon.

The government of the European Union, in its quest for an exascale supercomputer has targeted a goal of getting 50 gigaflops per watt (Intel also thinks this would work). In conversations with folks that design supercomputers, the thinking is that a conventional x86-based machine would require the equivalent of a power plant or two to run. That includes all the networking and other trimmings, but the bottom line is Adapteva’s chips deliver more flops per watt, and that’s a good thing.

It’s not just supercomputers though. Adapteva’s CEO Andreas Olofsson told me the company is only targeting computing extremes such as supercomputing and mobile phones because that’s where the power efficiency pain point is today. Because mobile phones run on batteries, and no one wants a smartphone that dies after 2 hours, vendors using ARM’s power efficient architecture have dominated the mobile sector. when Microsoft, adapted Windows to run on ARM, it spoke volumes about the need for power efficiency. Windows, is one of the most x86-oriented peices of software out there.

These shifts in usage profiles and the high demand for compute are creating opportunities for companies like Adapteva, so it’s not too far-fetched to wonder how long until that pain point hits conventional servers.

I often cover companies that are hoping that the combination of monolithic applications and a desire to reduce power consumption means that webscale and cloud vendors will embrace a new architecture. Companies such as Tilera, SeaMicro, Adapteva, Calxeda and others are all betting that the next gear Facebook or Amazon buys will be their hardware or contain their chips.

However, even in its state-of-the-art data center that’s optimized at the very server level to be energy-efficient, Facebook challenged the way servers and data centers are built but didn’t touch the silicon itself. So, clearly, the webscale world isn’t champing at the bit to replace the x86-based servers their applications are running on. SeaMicro even has shown charts showing that the CPU is only a third of the power associated with running a server, which means there’s still plenty of fat to trim. Of course, Seamicro is building a server that trims that non-CPU fat and runs Intel’s Atom chips.

However, the global demand for energy and the supply we currently have are reaching a point where it’s safe to conclude that power consumption will become a greater cost and constraint associated with operating data centers. And at some point building in cooler climates, hot and cold aisle containment, and even newly designed servers won’t be enough if the silicon itself is too hot.

So the question isn’t if, but when, server companies abandon the PC-style architecture. Perhaps Intel, AMD or Via will continue to tweak x86 silicon until it can perform more calculations using less power, or perhaps it will be time for Amazon or Microsoft Azure to go with ARM, Tilera — or even Adapteva.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Infrastructure Q2: Big data and PaaS gain more momentum
• Infrastructure Overview, Q2 2010
• What cell phones can teach us about energy efficiency

alt=''
border='0'
/>

GigaOM

Chris Marino vCider CEO

Structure LaunchPad startup vCider disguises packets like I disguise zucchini in my four-year-old’s chocolate muffins. Except vCider does it to help companies bridge between different cloud providers while providing a bit less latency and more security that one would with a virtual router.

VCider aims to help companies span clouds while keeping their security and compliance intact by installing software on a virtual machine instance in a cloud and then creating a mesh-like Layer 2 network that the administrator can control. But because it’s in an instance running at Layer 3 of the network, it’s akin to hiding Layer 2 packets inside the Layer 3 tunnels in order to allow for greater control of the network without adding a lot of latency. Using such clients inside each instance is becoming a more common practice as companies CloudPassage also experiment with that — only for security rather than bridging a network.

There are other companies such as Vyatta offering similar services to try to present a unified network view, but they use virtual routers that have their own limitations — namely greater latency and they can introduce a single point of failure.

VCider, which was founded in the fall of 2010, had raised a seed round of about 0,000 from undisclosed investors, according to CEO Chris Marino. It is looking for a Series A funding.

Related content from GigaOM Pro (subscription req’d):

• The Case for Increased M&A in 2011: Actions and Outlooks
• The Structure 50: The Top 50 Cloud Innovators
• In Q3, Big Data Meant Big Dollars

alt=''
border='0'
/>

GigaOM — Tech News, Analysis and Trends

In some cases, cloud computing is merely a means to avoid investing in “undifferentiated heavy lifting,” but when done right, it actually can be a source of significant competitive advantage. So says Zynga, at least, which highlighted its unique cloud infrastructure, as well as its advanced analytics efforts, as part of its core strengths in the S-1 statement it filed this morning.

According to the form, Zynga views its “scalable technology infrastructure” as a core strength, stating that, “We have created a scalable cloud-based server and network infrastructure that enables us to deliver games to millions of players simultaneously with high levels of performance and reliability.” In describing its cloud infrastructure as an important aspect of its business, Zynga’s S-1 says:

Our physical network infrastructure utilizes a mixture of our own datacenters and public cloud datacenters linked with high-speed networking. We utilize commodity hardware, and our architecture is designed for high availability and fault tolerance while accommodating the demands of social game play.

We have developed our architecture to work effectively in a flexible cloud environment that has a high degree of elasticity. For example, our automatic provisioning tools have enabled us to add up to 1,000 servers in a 24-hour period in response to game demand. We operate at a scale that routinely delivers more than one petabyte of content per day. We intend to invest in and use more of our own infrastructure going forward, which we believe will provide us with an even better cost profile and position us to further drive operating leverage.

Zynga has been touting its Z Cloud infrastructure for more than a year, which takes a reverse approach to the conventional approach to hybrid cloud computing. Whereas many analysts initially assumed companies would use private clouds as a gateway to public clouds, Zynga uses Amazon EC2 as a staging ground before ultimately moving games onto private cloud resources. Essentially, Amazon’s cloud lets Zynga scale elastically and determine average traffic load and other metrics, so that it can optimize its internal infrastructure for each game’s specific needs.

The goal of this strategy is efficiency: Zynga doesn’t have to invest in more resources than necessary upfront, nor does it have to worry about underprovisioning resources or otherwise inadequately configuring them when it brings games onto its private cloud. In many cases, private clouds can cost less than public clouds for applications with fairly stable usage patterns, and they help companies meet various requirements around security and compliance. Zynga uses Cloud.com for its private cloud infrastructure, as well as RightScale as a management layer that makes for a uniform experience in terms of managing both public and private resources.

As is the case with every leading web company, Zynga also highlights its big data strategy as a key differentiator. Describing its “sophisticated data analytics,” the S-1 notes, “The extensive engagement of our players provides over 15 terabytes of game data per day that we use to enhance our games by designing, testing and releasing new features on an ongoing basis. We believe that combining data analytics with creative game design enables us to create a superior player experience.”

Cloud computing and advanced analytics are double-edged swords, though. As Zynga’s S-1 acknowledges, relying on publicly hosted cloud computing resources makes it vulnerable to service outages like Amazon Web Services’ infamous April 2011 outage, which temporarily downed both FarmVille and CityVille. “If a particular game is unavailable when players attempt to access it or navigation through a game is slower than they expect, players may stop playing the game and may be less likely to return to the game as often, if at all,” the form states.

Relying on advanced infrastructures and analytics also means competing with companies such as Facebook, Google and others for employees skilled enough to keep Zynga’s operations on the cutting edge. Specifically, the company acknowledges, “game designers, product managers and engineers” are in high demand, making attracting and retaining them a resource-intensive process. In some cases, this has meant offering particularly attractive employees lucrative stock options, which could come back to bite the company. As it notes in the S-1, “[W]e expect that this [IPO] will create disparities in wealth among our employees, which may harm our culture and relations among employees.”

Image courtesy of Flickr user eschipul.

Related content from GigaOM Pro (subscription req’d):

• Infrastructure Q1: IaaS Comes Down to Earth; Big Data Takes Flight
• A field guide to cloud computing: current trends, future opportunities
• The Structure 50: The Top 50 Cloud Innovators

alt=''
border='0'
/>

GigaOM — Tech News, Analysis and Trends

Since the concept of “private cloud” was introduced, there have been efforts by certain people to prove it “wrong” or show that it doesn’t make sense when compared with the public cloud. This seems like a silly crusade, not because I’m a supporter of private cloud (which I am), but because both provide tremendous value if you actually understand the value that “cloud” has delivered to the industry.

“But Sinclair, cloud delivers value because of the economies of scale it brings through aggregation, etc., etc., and so on,” you say? Perhaps, but the value of cloud computing has much more to do with its definition in the abstract and less so to do with its availability in a public form factor.  James Urquhart recently wrote “Why definitions of cloud are creating ‘false’ debates,” where he hypothesizes (accurately so) that the difference of opinion is that some characterize “cloud” as a business model, while others as an operations model.

Clearly, when looked at from the business model point of view, the concept of cloud makes significant sense in a public fashion. But as an operations model — a model where resources are pooled together behind abstractions that dynamically manage applications and resources — it has significant positive implications in the enterprise. This might be easier to explain through an analogy of sorts.

Let’s suppose, for the purposes of this thought experiment, that the bus (the big automobile that carries lots of people) has yet to be invented. A politician notices the inefficiency of always using a car that fits no more than four people, particularly in the case where lots of people are going back and forth between two cities — the politician’s home city and a neighboring city. This politician decided that the cities should operate a municipal (or public) mass transit service to transport a significant number of people per trip between the two cities, for some small fee per person. The politician commissions the invention of the bus to transport 50-100 people at a time. The idea of offering this as a public service is powerful, and as the number of passengers grows, it starts to experience significant economies of scale.

All is well, until some suggest that the bus itself is useful in contexts outside of public transit. Schools want their own buses to pick up and drop off children; prisons determine buses are a good way to transport large numbers of prisoners; someone wants to start a luxury tour service via bus; and movie stars that hate to fly feel buying a bus is an effective way to travel along with their friends, family and staff.

The politician becomes angry, stating that all of those use cases are best satisfied via the public transport system she developed, and these “private” uses are “false mass transit services” because they could never reach the economies that the public service offers. Furthermore, she argues, these proponents of “private mass transit” are getting in their own way because the public transit system is not only cost effective, but safe and generally on time, and all of the constraints that these other use cases point to in usage of public transit are merely “excuses.” The fact of the matter is, to a bus-rider, riding in a bus provides the same end utility regardless of how the bus is provided – they get where they want without having to drive a car.

Does this seem awkward and familiar at the same time? It does to me. The problem is that the politician is lumping the invention of the bus — the technology necessary for public mass transit to work – and the public mass transit system itself into a single cohesive model, and taking the stance that the real marginal savings of public mass transit is the only economic output to take into consideration. Others have decoupled the bus from the public transit service, saying that although there is huge value in public transit, the bus itself adds so much value to a huge number of use cases (such as prisoner transport) that are ill suited for public mass transit because of constraints.

Without the bus, those “private” use cases are still using four-passenger cars for all their transport needs. However, the bus solves a significant number of problems relating to moving large numbers of people relatively efficiently without having to adopt public mass transit. Similar to Urquhart’s assessment, the problem in this bus analogy is that someone is focused on the public-transit business model while others are focusing on the operations-model efficiencies that the bus can bring to other use cases.

Confounding cloud computing from a service point of view with the technology that enables cloud services is terribly misguided. The fact of the matter is that tge technology behind cloud services is extremely valuable on its own, just like a bus is extremely value outside of the public mass transit context. Take Platform as a Service, for instance. PaaS provides a tremendous amount of agility through…

1. The pooling of resources (servers, load balancers, etc.) into a single abstract pool of resources
2. Automation of devops workflows, thereby increasing time to market
3. Utilization boosts (in multitenant environments)
4. Simplified management around previously complex topics (e.g., scaling out, etc.)

This value has nothing to do with economies of scale or outsourced IT, but has everything to do with a paradigm shift in the deployment and management of applications. If an organization chooses to layer a PaaS tier – a private PaaS - atop its own infrastructure, whether it be dozens, hundreds or thousands of servers, it will experience genuine value. The technology developed to supply PaaS is much more useful than just the fact that it’s offered as a service — it can drive a whole new era of efficiency as a layer in the private cloud stack, on top of an enterprise’s existing infrastructure.

This is why “false cloud” articles, like one by Phil Wainewright titled “Private cloud discredited, part 2,” disturb me. They are too myopic in terms of debate basis, focusing on economies of scale and not much else, and fail to separate the invention of cloud enabling software layers like PaaS (the bus) from their first use in the public cloud context (public mass transit). Just as throwing away the bus in any context other than public mass transit system makes little sense, dismissing the massive efficiencies achievable by deploying technologies like private PaaS would be crazy. As David Linthicum put it in a recent post: ”[M]any fail to accept there may be times when the architectural patterns of public clouds best serve the requirements of the business when implemented locally — in a private cloud.”

Sinclair Schuller is co-founder and CEO of Apprenda.

Image courtesy of Flickr user KB35.

Related content from GigaOM Pro (subscription req’d):

• Infrastructure Q1: IaaS Comes Down to Earth; Big Data Takes Flight
• Infrastructure Overview, Q2 2010
• Does Diversification Equal Domination in the Cloud Stack?

alt=''
border='0'
/>

GigaOM — Tech News, Analysis and Trends

Despite a lot of speculation lately about who’s winning the private-cloud race and what companies might be on the way out, it’s far too early to call the game in anyone’s favor. Private-cloud adoption is picking up, but it’s nowhere near ubiquitous yet, and there’s plenty of time for everyone still standing to make the moves they need to in order to keep competing.

Is Eucalyptus Dead in the Water?

There was a great debate last week on GigaOM Pro about the state of Eucalyptus Systems and whether its days as a purveyor of private-cloud software are numbered. Our Infrastructure curator, Paul Miller, suggested in his Weekly Update (sub req’d) that, after a few previous setbacks, Ubuntu’s decision to forgo further Eucalyptus support in favor of OpenStack in future versions of the Ubuntu Enterprise Cloud Linux operating system might be the death knell for Eucalyptus Systems’ private-cloud software.

Eucalyptus CEO Marten Mickos responded, stating that his company has had more than 25,000 downloads clouds based on its free version, and that, from his perspective, it’s stronger than ever. He also mentioned a number of big-name customers presently running production clouds based on Eucalyptus.

So did Cloudera CEO Mike Olson, who noted that his company is a very happy Eucalyptus customer. This is in part because of its Amazon Web Services API compatibility, and in part because of the qualities of the product and support for paid enterprise version of the Eucalyptus software. Olson’s most poignant comment might have been that “it’s way too early to nominate a single winner among the cloud abstraction layers. We need more years of experience before that happens.”

I’ve been critical of Eucalyptus chances, too — even suggesting at one point (sub req’d) that an acquisition might be all that could save it — but I’m starting to come around to Mickos’s point of view. His statements about Eucalyptus’s prospects are, of course, self-serving, but there’s no denying the numbers. Even if a majority of those 25,000 downloads clouds never amounted to anything substantial, some certainly did and will in the future. Yesterday, it announced European social-gaming company Plinga as a customer. And having large, publicly referenceable customers puts Eucalyptus ahead of many other private-cloud startups. Further, the company has continued to add personnel and expand globally, which probably aren’t signs of impending doom.

It’s a Broad — and Young — Market

But this isn’t just a discussion about Eucalyptus, it’s a discussion about the expectations for private clouds, in general. OpenStack has a lot of believers — and for good reason — but with the exceptions of Rackspace and Internap, there are no service providers that are known to be using the software for customer-facing services, and most private-cloud use cases appear only experimental at this point. As Mickos pointed out in his response to Miller’s analysis, Eucalyptus most commonly runs into VMware, Cloud.com, Abiquo and CA (with its 3Tera product, I presume) during customer engagements.

What this tells us is that, as Olson suggested, it’s too early to tell who (aside from VMware) will win in the private-cloud contest, even if some projects or companies have greater name recognition than do others. VMware is a household name and OpenStack is approaching that status, but Cloud.com, Abiquo and 3Tera are not. Yet, 3Tera was an early innovator in provisioning private clouds before CA bought it, Cloud.com is killing it with several very large customers under NDA, and even the relatively unknown Abiquo has a growing list of customers. Still under the radar, but not to be ignored, are startups such as Cloupia and Nimbula (which just became generally available in April), and large vendors with new cloud strategies, such as Red Hat, HP and Microsoft.

In fact, Nimbula Co-Founder and CEO Chris Pinkham insightfully mentioned to me during a recent conversation that the whole discussion about public clouds versus private clouds is just a debate over who owns the hardware. What customers really care about — or what they ultimately will really care about — he explained, is which product can best deliver a company’s services, regardless where the servers reside.

What he’s describing is the oft-mentioned but as of yet rarely implemented hybrid cloud. And despite some noteworthy efforts by pretty much everyone pitching private-cloud software, no one has this mastered yet. However, if you’ve looked at any cloud-adoption surveys lately, such as this one from the Open Group, you’ll find that hybrid clouds are all the rage among CIOs.

Yes, OpenStack has all the momentum right now, but it doesn’t obviate the need for other products, nor is it even a completed project. It’s a worthwhile exercise to handicap the private-cloud field, but with adoption still relatively low and looking to remain that way until enterprise-ready hybrid clouds become a reality, all we really have right now are actual customer wins and product roadmaps to determine who has the best chances. At this very-early point, a lot of products look promising, but there’s plenty of time for private-cloud pushers to distinguish themselves.

Image courtesy of Flickr user superwebdeveloper.

Related content from GigaOM Pro (subscription req’d):

• What Ubuntu’s Move to OpenStack Means for Eucalyptus
• Infrastructure 2011: The Real Cloud Computing Picture Will Emerge
• Infrastructure Q1: IaaS Comes Down to Earth; Big Data Takes Flight

alt='The cloud-optimized networks of tomorrow run on Brocade Ethernet fabrics today. Assess Your Cloud Readiness. Download Forrester Study »'
border='0' width='300' height='250'
/>

GigaOMGigaOM · Tech News, Analysis and Trends