Facebook set Open Compute in train a year ago to solve problems it was having trying to shoehorn the compute, storage and networking density it needed into the traditional server rack, a form factor its hardware master calls “blades gone bad.”

Blades supposedly go bad because of what OCP founding board member Andy Bechtolsheim calls “gratuitous differentiation” on the part of vendors and their lock-in-seeking proprietary designs that sacrifice interoperability.

Open Rack is about standardizing the electrical and mechanical interfaces in a rack and saving power. It is based on an airier OpenU (OU) that’s 48mm high instead of the old 44.5mm U so everything can breath and there can be better cable management and more efficient use of the space.

The Open Rack frame is still a familiar and reassuring 24-inch wide but the slot inside is 21-inch wide (though existing 19-inch equipment can still be accommodated). The wider bay can fit three two-socket motherboards or five 3.5-inch disk drives side-by-side.

It’s supposed to lower TCO by lengthening the compute components’ lifespan and reducing industrial waste. OCP imagines that whole servers won’t have to be replaced every two-and-a-half years, just the components, some of which could be good for 10 years. And when Intel or AMD come out with a new dingus just the processor gets upgraded.

Servers won’t have their own power supplies anymore either. They’ll plug into a 12V bus bar at the back of the rack that in turn connects to power shelves in each rack. Bye-bye cables.

Open Rack has three sizes: a triple rack, a single rack and a half rack. Facebook fancies the roomier triple rack.

HP and Dell have new so-called “clean sheet” server and storage designs compatible with Open Rack specification. HP’s is called Project Coyote and Dell’s is called Zeus.

Coyote uses two Xeon E5 processors and is supposed to represent a power savings of up to 50% compared to standard racks. Dell’s widgetry mixes different kinds of servers and storage.

HP, the world’s largest server maker, and Quanta, the world’s largest contract manufacturer, have just joined OCP. So have AMD, Fidelity Investments, Salesforce.com, VMware, Canonical, Avnet, Alibaba, Supermicro, Cloudscaling and Tencent. IBM is still being stand-offish.

HP, Quanta and Tencent have joined the OCP Incubation Committee, which reviews proposed projects to determine whether they should receive official OCP support. The committee is already considering a Facebook design for a “vanity-free” storage server code-named Knox and what are supposed highly efficient motherboard designs for financial services companies from AMD and Intel. AMD’s is called Roadrunner and Intel’s is called Decathlete.

Frank Frankovsky, Facebook’s VP of hardware design and supply chain and OPC chairman and president, says they’re also “mapping out a convergence between Open Rack and Project Scorpio, a similar spec under development by the Chinese Internet giants Tencent and Baidu. They expect the two specs to merge next year.

VMware is promising to certify its vSphere virtualization platform to run on OCP gear, and DDN says it will do the same with its WOS storage system. Canonical means to certify its widgetry for OCP servers before the designs are released.

And there’s now an OCP Solutions Provider program that’s supposed to help companies “sell and consume technology based on Open Compute Project designs.” Hyve, ZT Systems, Avnet and new business units at Quanta and Wistron called QCT and Wiwynn, respectively have been launched to sell directly to consumers and are in line for Solutions Provider status.

Frankovsky said in a blog posting that “The momentum that has gathered behind the project – especially in the last six months – has been nothing short of amazing.”

Everyone agrees it has Amazon, particularly, and VMware, to a certain extent, in mind. However, what’ll probably end up happening is that Red Hat, which has a prominent part in the open source OpenStack project that Rackspace started, becomes the “Linux of the cloud” because it’s got all the pieces, or thinks it does, but that’s another story.

Anyway, Rackspace is inching out with a production-ready OpenStack cloud based on Essex, the fifth and best-yet release of the open source cloud platform put in train by Rackspace and NASA in the summer of 2010.

Rackspace CEO Louis Napier told a New York Times blog that he expects to have all his customers, by then perhaps 200,000 businesses, on some or all of an OpenStack system by summer.

The fight is supposed to come down to a dual between the proprietary Amazon APIs, now lauded as the de facto standard of public clouds, and the still immature but open source CloudStack APIs.

Rackspace says come May 1, in roughly two weeks time, it will begin providing customers with default access to widgetry that it’s now got in “limited availability.”

That includes:

* Cloud Servers, the Essex-based EC2-like compute piece of OpenStack, a k a Nova, accessible through the new programmable OpenStack API for switching between OpenStack clouds or a new intuitive control panel.

* A built-from-the-ground-up graphical Control Panel that allows server tagging to discriminate between production and development servers so they can be controlled in concert and has multi-region capabilities.

Rackspace says “limited availability” means customers can sign up now, the widgetry is reportedly production workload-ready, there are unspecified SLAs, 24×7 support and regular billing. It seems it could take Rackspace a couple, few months to ensure a smooth ramp-up but the move is supposed to be imperceptible.

Rackspace, which has already got the S3-like Cloud Files storage, a k a Swift, will move all of its public, private and hybrid cloud to this widgetry.

It’s also got stuff in “early access” defined as “production workload-ready but with limited support available, no service commitments and no billing.

They include:

* OpenStack Cloud Databases (Project Red Dwarf) with API access to a massively scalable, highly available MySQL database with redundant SAN storage for high performance and automated management. Figure on Microsoft SQL Server and other databases too but maybe not from Rackspace. Amazon, of course, has the MySQL-derived Relational Data Service (RDS).

* Single-view Cloud Monitoring of the infrastructure and applications based on Rackspace’s acquisition of Cloudkick.

Lastly it’s got early versions of products in “preview” looking for testers namely:

* OpenStack Cloud Block Storage, like Amazon’s Elastic Block Storage (EBS), but offering either solid state or lower-cost disk storage.

* Cloud Networks, software-defined virtual networks for managing logically abstracted network services programmatically. IDC says Rackspace’s Cloud Networks is “going to eliminate some of the hesitation businesses have around cloud adoption.” Thank you Cisco et al.

Rudimentary pricing will remain the same starting at $0.015 cents an hour for a Linux virtual server with 10GB of disk space and 256MB of RAM and $0.08 an hour for Windows.

Rackspace still has to say what the database, storage and networking will ultimately run.

HP, a chief OpenStack acolyte, won’t have a beta take on the Essex platform memorialized in its HP Cloud Services until May 10 with no estimates, as of last week, on when it could have a production public cloud.

There will be OpenStack fragmentation, observers prophecy. HP won’t return all the distinguishing tweaks it makes to the community.

Rackspace says on its web site that it’s got more than 170,000 businesses and 60% of the Fortune 100 as customers.

See www.rackspace.com/nextgen.

First, it dumped its OpenStack distribution, Project Olympus, in the garbage.

Then it took CloudStack, an asset it spent somewhere between $200 million and $250 million to acquire nine months ago, and turned it over lock, stock and every last shred of legally unfettered proprietary code to the Apache Software Foundation (ASF), which has nurtured Hadoop, Hive and Cassandra/

ASF will incubate CloudStack as an open source rival to OpenStack and a foil to proprietary VMware, Citrix’ hereditary enemy.

And then it loudly proclaimed to anybody who would listen that Amazon’s proprietary APIs are the de facto standard and that CloudStack would fully embrace Amazon compatibility.

The move puts CloudStack in competition with Eucalyptus Systems, the open source private cloud peddler that just cut a deal with Amazon that was years in the making to guarantee that its Amazon EC2 and S3 APIs are in prefect pitch with Amazon’s APIs so that companies can have a private cloud that gracefully migrates workloads to Amazon.

Considering which side its probably-now-billion-dollar bread is buttered on, Amazon doesn’t like private clouds, but by bringing Eucalyptus into the fold it has conceded that it has to tolerate hybrid clouds to get more enterprise business. It’s betting that over time the enterprise will get so accustomed to Amazon they’ll abandon their currently safer private clouds.

The unique Eucalyptus licensing deal with Amazon, a boon to both, might not be exclusive, but it also might not be something Amazon is ready to duplicate right quick, leaving CloudStack adherents to worry about broken interoperability and the possible infringement of its clones.

Undaunted, the general manager of Citrix’ Cloud Platforms Group Sameer Dholakia says, “Amazon has invented and created this market, and with what is projected to be $1 billion in ecosystem and customer revenue attached to Amazon cloud, we believe the winning cloud platform will have to have a high degree of interoperability with Amazon.” (No vendor-neutral agonizing or carping over who controls the standard or how it’s developed there though there’s at least one untrusting call for Amazon to define an Amazon Community Process.)

Citrix, which was among the first of what are now 150 companies to rally to the open source Rackspace-NASA OpenStack project two years ago, said it can’t wait around another year or two for the immature OpenStack to get debugged, stable and feature-complete like vCloud when CloudStack – which was supposed to be integrated into OpenStack – is production-ready and scalable.

OpenStack, it said, wasn’t responding to customer needs.

The OpenStack community of, shall we say, um, “free spirits” ultimately rejected the idea of integrating CloudStack completely, and evidently Citrix had problems with the politics of OpenStack, which has vendors like Dell, Cisco and HP (and maybe IBM) behind it, while service providers tend to fancy CloudStack and have deployed it for real, which is more than OpenStack can say.

Citrix also complained that, although OpenStack pays lip service to the Amazon APIs, it prefers its own APIs, something that older and wiser heads close to OpenStack say will ultimately have to change, attributing it to the naïveté of the purists in the OpenStack community.

It’s also possible that Citrix really, really, really wants to sell more XenServer to justify the $500 million or so it blew buying XenSource a few years ago.

In the interests of XenServer, Citrix is going to hang on the periphery of the KVM-leaning OpenStack and make contributions to see its hypervisor supported.

Meanwhile, CloudStack 3.0 is going to trade in its GPL 3 open source license for the friendlier Apache 2.0 license, making it the first cloud platform to sport an Apache license – which should prove attractive to both companies and the many developers in the Apache community.

Citrix also gets to say that CloudStack is sheltering in a legitimate, neutral foundation not an interests-controlled one like OpenStack’s.

CloudStack has 50-odd software merchants and SPs – like BT, Engine Yard, GoDaddy and Korea Telecom – supporting it and upwards of a hundred clouds is production, some reportedly with tens of thousands of servers that are reportedly generating more than a billion dollars for their owners. It’s also got 30,000 community members and reportedly thousands of certified apps. It also picked up a big OpenStack defector, NTT DoCoMo, the giant Japanese telco.

While hoping for and positioning itself for broad adoption, sources say Citrix’ real reason for contributing CloudStack to Apache was to retain this following, which obviously might be uncomfortable with a purely Citrix-led defection from OpenStack despite the clear superiority of the compute part of CloudStack.

Amusingly, Citrix allows that the Java-based CloudStack might someday suck up some mature parts of the Python-based OpenStack.

What is probably going to happen is that CloudStack winds up a hybrid mixture of its own compute and OpenStack’s prized S3-mimicking “Swift” storage contributed by Rackspace.

At least the smart money among the CloudStack contingent is expected to do that, leaving folks like the perennially benighted and ponderous HP up a creek as usual.

Anyway, as a way to get some of its money back, Citrix said it will continue to offer a commercially supported Apache CloudStack distribution tuned to XenServer (which, by the way, offers indemnification, hmmm) and spend millions of dollars on engineering and marketing. It will have to to be heard over the OpenStack hoopla.

In its attempt to position CloudStack to become the “de facto of cloud computing platforms,” Citrix said it would become a platinum sponsor of the Apache Software Foundation like Facebook, Google, Microsoft and Yahoo.

Oh, yeah, Eucalyptus’ IaaS, by the way, has 100 paying customers and, say, about 25,000 downloads 10% of which are regarded as serious.

Gartner called the Citrix defection a “bombshell,” the beginning of “the war for open source clouds.”

See CloudStack.org community.

These new commitments are supposed to improve the tools and techniques needed to access, organize and glean discoveries from huge volumes of digital data.

Dr. John Holdren, director of the White House Office of Science and Technology Policy, said, “In the same way that past federal investments in information technology R&D led to dramatic advances in supercomputing and the creation of the Internet, the initiative we are launching today promises to transform our ability to use Big Data for scientific discovery, environmental and biomedical research, education and national security.”

It’s seen as being that important.

The major initiative is supposed advance state-of-the-art core technologies, apply them to accelerate the pace of discovery in science and engineering as well as transform teaching and learning, and expand the workforce needed to develop and use Big Data technologies.

It’s a response to recommendations by the President’s Council of Advisors on Science and Technology, which last year concluded that the federal government was under-investing in Big Data technologies.

As a result, the National Science Foundation (NSF) and the National Institutes of Health (NIH) will be implementing a long-term strategy that includes new methods to derive knowledge from data; infrastructure to manage, curate and serve data to communities; and new approaches to education and workforce development.

As a start, NSF will be funding a $10 million Expeditions in Computing project based at Berkeley to integrate machine learning, cloud computing and crowd sourcing.

It will also provide the first round of grants to support EarthCube, a system that lets geoscientists access, analyze and share information about the planet, issue a $2 million award for a research training group for undergraduates to use graphical and visualization techniques for complex data and provide $1.4 million to support a focused research group of statisticians and biologists to determine protein structures and biological pathways.

NIH is particularly interested in imaging, molecular, cellular, electrophysiological, chemical, behavioral, epidemiological, clinical and other data sets related to health and disease.

It said the world’s largest set of data on human genetic variation – produced by the international 1000 Genomes Project – is now available on Amazon’s cloud. At 200TB – the equivalent of 16 million file cabinets filled with text, or more than 30,000 standard DVDs – the current 1000 Genomes Project data set, derived from 1,700 people, is a prime example of Big Data.

AWS is storing the 1000 Genomes Project on S3 and in Amazon Elastic Block Store (EBS) as a publicly available data set for free; researchers only will pay for the EC2 and Elastic MapReduce (EMR) services they use for disease research. They used to have to download publicly available datasets from government data centers to their own systems, or have the data physically shipped to them on disks. The current aim of the project is to sequence 2,600 individuals from 26 populations around the world. (See http://aws.amazon.com/1000genomes.)

The Defense Department will be investing around $250 million a year (with $60 million available for new research projects) in a series of programs that use Big Data in new ways to bring together sensing, perception and decision support to make autonomous systems that can maneuver and make decisions on their own.

The agency also wants a 100-fold increase in the ability of analysts to extract information from texts in any language, and a similar increase in the number of objects, activities and events an analyst can observe.

DARPA, the Defense Advanced Research Projects Agency, is beginning an XDATA program that will invest about $25 million a year for four years to develop computational techniques and software tools for analyzing large volumes of data, both semi-structured (tabular, relational, categorical, metadata) and unstructured (text documents, message traffic).

That means developing scalable algorithms for processing imperfect data in distributed data stores and creating effective human-computer interaction tools to facilitate rapidly customizable visual reasoning for diverse missions.

The XDATA program will employ open source toolkits for software development so users can process large volumes of data in timelines “commensurate with mission workflows of targeted defense applications.”

The Energy Department will kick in $25 million in funding to establish a Scalable Data Management, Analysis and Visualization (SDAV) Institute under Lawrence Berkeley National Laboratory.

It’s supposed to bring together the expertise of six national laboratories and seven universities to develop new tools to help scientists manage and visualize data on the agency’s supercomputers to streamline the processes that lead to discoveries made by scientists using the agency’s research facilities. It said new tools are needed since the simulations running on its supercomputers have increased in size and complexity.

Lastly, the US Geological Survey will incubate Big Data projects that address issues such as species response to climate change, earthquake recurrence rates and the next generation of ecological indicators.

The old-line Big Data house already has an alliance with Cloudera, the first-to-market Hadoop commercializer, and if Teradata customers are absolutely positive they want Cloudera’s Hadoop distribution they can get it from Teradata.

However, Teradata’s new non-exclusive alliance with Hortonworks, which can out-credential its rivals, will give it more of an air of authenticity since a lot of the guys involved in getting Hortonworks off the ground built Hadoop for Yahoo, the code’s originator, and the code that now underpins Facebook, Twitter and eBay.

Anyway, Cloudera’s also in bed with Oracle.

The industry’s newest pair is supposed to do joint R&D, support and marketing and deliver integrated enterprise architectures that tackle all aspects of Big Data processing including a set of reference architectures for data-crunching Hadoop clusters.

Teradata sees volumes of raw data being poured into Hadoop then passed to the Aster Data SQL-MapReduce analytics widgetry Teradata bought last year for refining and moving on to a Teradata Database where normal people, well, relatively normal people can consume it and draw inferences.

Having being collaborating for a while, they already have connectors for Hadoop to Aster and Hadoop to Teradata’s Database as well as back again so Hadoop can store data.

The pair said “businesses will be able to quickly load and refine multi-structured data” – such as weblogs, text and customer interaction data – “some of which is being discarded today, for discovery and analytics.”

Teradata takes the position that Hadoop, over and above being crotchety to set up, deploy and manage, has confused everybody.

Tasso Argyros, co-president of what is now Teradata Aster, blogged that “It is usually not clear what use cases apply to traditional technologies versus new; how to reconcile existing technologies with new investments; and what type of projects will provide the highest ROI versus a long and painful failure. Teradata and Teradata Aster…want to earn the trust to advise our customers on how to use complementary solutions, like Hadoop – and make sure that the total solution works and reliably succeeds in tackling big business problems.”

He says, “Our goal is to NOT let even one enterprise customer fail with a Big Data project. We have enough collective experience to guide customers to avoid failed projects and traps.”

This is where Hortonworks comes in. It’s supposed to explain the role of Hadoop in the data center, and how Hadoop technologies can bring value to the enterprise.

The New York Times observes that the alliance “to the extent that it adds credibility and exposure to both Teradata and Hortonworks it is likely to lead to a thinning out of the number of different players, and squelch the likelihood that different and incompatible types of Hadoop will evolve” since enterprise customers want a stable and open Hadoop.

Hortonworks has previously projected that in five years or less more than half the world’s data will be stored in Apache Hadoop.

By the way, Hortonworks, which has designs on being the Hadoop market leader – and the tie-up with Teradata should broaden its customer footprint – confirmed that its initial CEO Eric Baldeschwieler, the former VP of software engineering for the Hadoop team at Yahoo, has been replaced by COO Rob Bearden, previously COO of both SpringSource and JBoss, two successful (and ultimately acquired) commercial open source companies, who was handling Hortonworks’ business side.

He’s supposed to be good at making open source operations ultimately turn a buck.

Baldeschwieler is now CTO and Baldeschwieler’s hires, VP of engineering Mark Himelstein and VP of customer support Marko Nicosia, are reportedly gone.

According to GigaOm the start-up has collected $50 million in VC backing, a number we couldn’t get it to confirm.

Microsoft has got Hortonworks helping with its Hadoop-relieved SQL Server 2012 and Hadoop on Azure.

And Porticor delivers its data encryption solution to IaaS and PaaS users through the cloud in minutes. Fancy that.

It’s supposed to solve the biggest challenge for data encryption in the cloud – storing keys.

It promises that a user’s data encryption key will never be exposed and that it can deliver data security across virtual disks, databases, distributed storage and file systems.

All this wonderfulness, called the Porticor Virtual Private Data (VPD) System, a combination of the start-up’s Virtual Appliance and Virtual Key Management Service, comes complements of its patent-pending homomorphic split-key encryption technology, which is supposed to increase security by an order of magnitude through hosted key management.

It’s supposed to be the industry’s first solution to combine data encryption with patented key management in defense of critical data in public, private and hybrid cloud environments.

Users can supposedly kiss good-bye traditional data security solutions that require costly software licenses and create operational overhead.

Porticor’s widgetry is a cost-effective virtual appliance that requires no encryption or key management experience to encrypt customers’ entire data layer with a proven AES 256-bit encryption algorithm.

The start-up expects the breakthrough to mitigate concerns about adopting the cloud.

VPD may best be compared to a Swiss bank. Entry to a Swiss lockbox needs two keys: one in possession of the bank, the other in possession of the owner.

That’s what Porticor does. It takes a patented split-key approach. Each data object, such as a disk or file, is encrypted with a unique key that’s split in two: a master key and a specific key. The master key is common to all data objects of one application and remains the possession of the application owner and is unknown to Porticor.

The second, “specific” key is different for each data object and is stored by the Porticor Virtual Key Management Service. As the application accesses the data store, Porticor uses both keys to dynamically encrypt and decrypt the data.

When the master key is in the cloud, it is said to be homomorphically encrypted – even when in use – and can never be seen in the cloud.

This mathematical technique lets Porticor do key-splitting and key-joining without knowing the key. It only knows the encrypted form of the keys.

By leaving one encryption with the customer Porticor differs from cloud encryption solutions that put customer encryption keys in the hands of a security vendor or cloud provider.

The widgetry complies with SOX, HIPAA, PCI DDS and GLBA, and reportedly solves the issues raised by the EU Data Protection and the US Patriot Act.

Porticor’s VPD system is available now. No download is required. It is deployed in the cloud and managed from Porticor’s customer portal. Pricing starts at $27.50 a month per Porticor Virtual Appliance for testing and very small production environments. The largest production environments run $411 a month. It comes with or without configuration for those who like to muck around but then setup takes longer.

It supports Amazon’s Virtual Private Cloud (VPC), where Porticor’s virtual appliance would logically go alongside the user’s application servers so the data never leaves the VPC unencrypted. Users can also snapshot their EBS disks, which are also encrypted.

Red Hat is reportedly offering the Porticor widgetry in its Cloud Foundation.

Porticor, which got started in 2010, is backed by $1 million A round from Glilot Capital. CEO Gilad Parann-Nissany said the company is working with Fortune 1000s.

The development is called the SeaMicro SM10000-XE. Needless to say, it’s the first fabric-based Xeon microserver ever made.

It’s also supposed to be the most energy-efficient, highest-density, highest-bandwidth Xeon server now available, period.

A single SM10000-XE replaces 32 dual-socket servers, but draws half the power and takes up a third the space without any changes to operating systems, applications or management tools.

It eliminates layers of Ethernet switches, server management devices and expensive load balancers.

Because it’s Xeon-based it can run heavyweight scale out workloads while the Atom-based SM10000-HD is meant for the highly parallel workloads found in the web tier. That means that microservers can capture a bigger piece of the data center. SeaMicro says the microserver will go mainstream running Apache Hadoop, real-time analytics, Java apps, PHP, Memcached and NoSQL.

The widget is built around the Xeon E3-1260L based on Sandy Bridge architecture and Samsung’s new low-power 1.35V Green DDR3 DRAM memory. Since it’s breakthrough, Intel and Samsung were both at its launch Tuesday when SeaMicro CEO Andrew Feldman announced it was shipping in volume.

Microservers are all about density, space and power. The Xeon E3-1260L has a 45W TDP envelope, providing 30% better performance per watt than processors from Intel’s previous generation, and SeaMicro has invented a new technology called TIO, short for Turn It Off, that lets it power-optimize the parts by turning off unnecessary CPU and chipset functions.

The chip’s four cores run at 2.4GHz CPU core. The clock can be throttled to 3.3GHz with Intel’s Turbo Boost Technology 2.0 and it’ll support eight threads with Intel Hyper-Threading Technology.

The SM10000-XE contains 64 CPUs for 256 2.4GHz cores in a 10U or 1,024 cores in a standard rack. It delivers 10 gigabits of bandwidth to each quad-core processor, which SeaMicro says sets a high watermark for bandwidth per unit compute.

It supports up to 32GB of DRAM per socket for a system total of 2.04TB.

The processors are tied together by SeaMicro’s Freedom Supercompute Fabric, which delivers 12 times the bandwidth per unit compute of a traditional server.

The XE also supports up to 16 10-gigabit Ethernet uplinks or 64 one-gigE uplinks. And unlike other microservers, it can support up to 64 SATA hard disks or solid-state disks without reducing computational density.

SeaMicro’s boards are Spartan in their minimalism consisting of only the processor, Samsung’s DDR3 and SeaMicro’s new Fabric ASIC, reportedly the industry’s first second-generation fabric ASIC, capable of supporting both large cores like the Xeon’s and small cores like the Atom’s.

SeaMicro figures eventually it’ll be able to design systems using Atoms and Xeons together in the same enclosure.

The XE’s austerity derives from SeaMicro’s Input/Output Virtualization Technology which reduces the component count, shrinks the motherboard and reduces power, cost and space.

The Freedom Supercompute Fabric is built of multiple Freedom ASICs working together, creating a 1.28 terabits a-second fabric that ties together 64 of the power-optimized mini-motherboards at low latency.

Samsung’s small high-density 30nm Green DDR3 is supposed to achieve more than a 70% power savings over 1Gb 1.5V 50nm class DDR3, improving SeaMicro’s TCO.

The SM10000-XE lists for $138,000 for a base configuration.

CompSec, which works with the US intelligence community, has adopted the widgetry in mission-critical applications to deliver huge amounts of compute power to remote and hard-to-access locations. It says it “greatly reduces response times, provides faster and more in-depth analysis, and helps to advance their mission.”

Mozilla is also on board. Matthew Zeier, director of IT infrastructure and operations, says that between the Atom boxes and the Xeon boxes “SeaMicro is able to meet the computational needs of our entire data center while driving down operating expenses by saving us power and space. We are a very happy customer.”

Intel has previously acknowledged that microservers could represent 10% of the server market by 2016. Its data center unit general manager Jason Waxman explained Tuesday that that forecast included both Atom and Xeon widgetry. Intel may be somewhat conservative considering the Xeon chip SeaMicro is using costs less than $300.

SeaMicro expects the next Atom generation to move chipset functions into the SoC, which will save more power and give it more real estate for processors.

SeaMicro rival Calxeda is building ARM chips into microservers.

In their place Egnyte (given the silly way we spell things you’re supposed to say ignite) proposes its HybridCloud, a next-generation storage, file-sharing and backup scheme originally targeted at SMBs that lately – at least in the last three quarters – has been attracting large accounts, including 30 departments in the Fortune 1000.

It’s thought to resolve a psychological barrier to cloud adoption by reassuringly keeping copies of what’s in the cloud on-premise. That’s obviously why they call it HybridCloud. Companies aren’t supposed to feel they’re losing control of their data.

Large accounts are also comfortable using FTP, which can move large amounts of data up and down. That’s why Egnyte has come up with the unified FTP and single sign-on capabilities that are supposed make separate FTP servers obsolete.

The start-up has 5,000 active trials going on all the time and says a third of them are FTP-related. According to CEO Vineet Jain “We know that two major pain points for IT are FTP and the ability to allow users to have a single sign-on when accessing services.”

With an FTP (or FTPES) transfer option integrated into the HybridCloud file server, users have a familiar way to securely upload large files directly into shared and private folders that are accessible from any device.

IT managers can set up and administer users directly from their Egnyte admin panel, using existing permission structures and eliminating those infernal special passwords.

For more complex or batch processes, scripting is supported, and all the files are accessible online or offline via Egnyte’s Office Local Cloud or Personal Local Cloud – the widgetry that keeps copies of files back in the office.

To streamline the process of using and maintaining user accounts while preserving security, Egnyte now supports the SAML protocol. End users can sign into their network once and automatically use Egnyte’s services without multiple passwords. They can also use Citrix Netscaler Cloud Gateway and OneLogin to go directly to Egnyte. IT managers can ensure Active Directory credentials are maintained within the firewall and integrate with existing SAML implementations, maintaining the critical extra security levels required in today’s cloud-infused world.

The start-up also argues – somewhat undeniably – that with cloud computing file server complexity and maintenance issues have increased dramatically and there are security issues galore.

It hired Forrester which did a study and found that 41% of employees at SMBs are using unauthorized cloud services in the workplace.

It says that means IT managers have to contend with a variety of consumer-grade products (presumably from its cloud storage competitors like Mozy and Box) that erode security, block IT from understanding what users are doing with critical company data and where they are storing it, ultimately impeding their company’s ability to collaborate and work effectively.

This time through Egnyte has expanded its syncing capabilities to include granular sub-folder level controls. The company, which figures nobody else has this kind of granularity – well, it did take nine months to develop – says serious cloud file-sharing means going beyond an easy-to-use interface and requires enterprise-grade security, IT administration tools, full auditing capabilities and syncing beyond a PC and Mac to include mobile devices, network attached storage (NAS) devices, storage area networks (SANs) and virtual servers.

Allowing cross-platform sync across a spectrum of devices is supposed to be the only way to make all employees, regardless of where they are, feel like they’re working from a single location.

So the widgetry now includes object-level integration with salesforce.com so users can collaborate with others regardless of whether or not they’re salesforce users, avoiding salesforce license fees. Access to salesforce files includes using iPad, iPhone, Android or some other mobile device. Salesforce.com users can work with files of any size, access virtually unlimited storage and share files directly within the salesforce.com interface. Team members who don’t have salesforce.com licenses can access the shared folders associated with salesforce objects through Egnyte.

Egnyte’s corporate plan covers 30 power users and 600 standard users and costs $228 a month when paid annually. That works out to $7.60 a power user a month.

Egnyte currently has a half-million users at what it says are thousands of companies representing over a billion shared files, taking advantage of a technology that offers the speed and security of local storage with the flexible accessibility of the cloud.

Egnyte is backed by $17 million from Kleiner Perkins, Floodgate Fund and Polaris Venture Partners.

The company has three data centers rented from Equinox in California, North Carolina and Amsterdam.