The use of solid state-based storage systems is rapidly increasing. So far, solid state technology has been deployed to accelerate applications in specific environments. Successes have been demonstrated for increasing the number of transactions a system can perform, the number of virtual machines per physical server (commonly referred to as virtual machine density), and the amount of virtual desktops supported by a storage system.
The continuing advance of all-solid state storage systems is leading to strategies where primary storage — defined as the most active storage of information for applications — will be all solid state. Planning for all-solid state primary storage requires the fabric infrastructure to be considered as a critical element in delivering the maximum value from solid state technology.
Solid state storage systems have much lower latency than systems design for spinning disks. The latency is measured in microseconds and the systems can sustain a much greater number of operations per second. Solid state technology is really a memory technology and using low-level disk-based access protocols may not be optimum. Faster or more streamlined protocols may reduce overhead and contribute to improvement in reducing latency.
So, what fabric interconnect is best? Most arguments about deploying new fabrics or making an infrastructure change have been based on cost. If the fabric interconnect technology requires additional hardware components to reduce latency, maybe the fabric technology with the lowest overhead or latency is the most economical. Economic valuation needs to be based on the increase in the efficiency of the system. Solid state systems can store and retrieve information faster, allowing applications to generate more transactions per second and deliver more value from the investment in servers and other hardware. The fabric choice to support solid state systems is a much bigger economic potential than the cost of the fabric or administration alone.
Low latency interfaces used today such as Fibre Channel or InfiniBand might deliver the greatest value when economic measures are used. Or, maybe another fabric or variation could evolve and create a disruption in the storage infrastructure. The economic value could make a compelling reason to make a transition.
Infrastructure for storage has always been a slow transition and that trend is expected to continue. But the efficiency and economic value delivered from solid state storage may accelerate a change of some type. The fabric decision will be based on how to enable applications to do more work and provide faster access to information. Initially, primary storage will be the focus for a fabric that can maximize value. Secondary storage may not be as demanding, but a common fabric may be preferred over a specialized fabric. Vendors will promote what they have now as a practical matter but will also look for the competitive advantage in delivering the economic value with future offerings. It just may take some time to evolve.
(Randy Kerns is Senior Strategist at Evaluator Group, an IT analyst firm).
Alex Bouzari, CEO of HPC storage vendor DataDirect Networks (DDN), said that will change in 2013. He expects big data to come into the mainstream and drag HPC with it. He expects DDN to come for the ride into the mainstream after more than a decade of handling big data needs before anybody called it big data.
“High performance computing has come of age,” Bouzari said, “and it’s now called big data.
“Big data is really the democratization of high performance computing. What was limited to a small number of extreme requirements has become commonplace. Now we’re seeing big data and high growth data across markets – the web, cloud and commercial high performance segments.”
It could be wishful thinking on his part, but Bouzari said big data requirements have already spilled into the enterprise, especially financial services, healthcare and manufacturing. He expects that to result in more cloud storage implementations as companies struggle to store and analyze data for their businesses.
That could be a boon to object-storage systems built for cloud scale, such as DDN’s Web Object Scaler (WOS) system. Bouzari also sees a need to make Hadoop work better with storage built for big data.
“Customers say ‘we love what Hadoop can do for us, but we need it as a product that can solve a business problem,’” Bouzari said. “Customers are looking to maximize the performance of Hadoop. We can greatly accelerate Hadoop.”
Bouzari also sees flash playing a role in big data, although until now the price of solid state storage has made it cost benefit for smaller data sets that need high performance. Bouzari said he is starting to see that change among DDN’s customer base.
“All-flash storage for high performance computing is one of those things that seemed to make a lot of sense, but proved cost prohibitive in many environments,” he said. “We’re did some all-solid state deployments [in 2012] but they were typically deployed as part of much larger IT infrastructures. As the cost of non-volatile memory continues to decrease, the ability to use it to serve huge amounts of content will make solid-state more attractive.”
Scality and Silicon Graphics Inc. (SGI) have signed an OEM deal that combines object storage software with high-performance commodity hardware. It provides a scale-out computing offering for organizations looking to build out systems that can store large amounts of unstructured data.
The deal includes the Scality Ring Organic software that resides on distributed, independent nodes and the SGI Modular InfiniteStorage Server (SGI MIS Server). The server delivers high density by packaging two motherboards together with up to 72 3.2 inch SAS or SATA drives in a single 4 U chassis. Twenty nodes in a 19 inch rack can hold 2.8 Petabytes.
“I think this move is symbolic because partners need each other pull together solutions,” said Ashish Nadkarni, IDC’s research director for storage systems. “SGI did not have native object storage. They (probably) had customers that said they needed object storage but SGI didn’t have it. They would be losing out on this part of the market, so they looked to partner with a hardware vendor.”
Floyd Christofferson, SGI’s director of storage product marketing, said object storage is software-focused “but really software has to run on hardware.”
Nadkarni said the SGI has expertise in high performance computing in large environments with massive storage systems. This gives Scality an opportunity to sell into SGI’s existing customer base.
Scality Ring presents a unified data interface delivering local and remote object and file access. The configuration can offer geo-redundancy and auto-tiering.
“This is a ring architecture that is designed to be scalable. You can keep adding hardware and capacity without compromising performance,” Nadkarni said.
There are other examples of the object storage ecosystem expanding. Cleversafe has formally announced a Solution Network partnership to boost sales of its object-based storage products. The partners include Symantec, CommVault, Cloudera, Riverbed, Cisco, VMware, Intel and Panzura.
“We are focused on a self-certification program that allows vendors to run tests [with Cleversafe software] to ensure the technology works with object storage,” said Russ Kennedy, Cleversafe’s vice president of product marketing.
Information Technology (IT) organizations must continue to innovate and advance their capabilities to deliver services more economically. Without making progress, competitive pressures will overtake IT with the following results:
• The potential to be outsourced will become greater with the promises (irrespective of the reality) of being less costly with more rapid reaction to changing demands.
• Internal groups demanding immediate resources will start looking for alternatives leading to rogue IT operations or rogue use of cloud services. Undoing these decisions can be difficult, both financially and without acrimony.
• Falling behind in IT can hold the company back from maintaining or gaining in a competitive environment.
Several storage technologies can be major contributor to innovation that advances IT. Technologies such as solid state storage for acceleration of access to information can make a major difference in the amount of processing an application can do. Embedded storage system capabilities for data protection using snapshots and remote replication continue to improve with new developments from vendors. Capitalizing on these improvements improve the operational environment and add to moving IT capabilities forward.
But storing and retrieving data with the performance necessary to serve applications is just the first part to consider. How efficiently the information can be managed is the real test of technology advancement.
Information is data with context where the context indicates the value, usage, and form necessary for understanding and using the information. Storage systems that can manage the value and automate the movement to different economic-basis storage and handle data protection provide major advancements in efficiency and ultimately result in greater competitiveness.
IT must make investments in storage technology that advances the capabilities of handling information. Each investment has a return on investment (ROI) payback that can be calculated, and an improvement in total cost of ownership (TCO) over the longer term. More importantly, the investments advance IT capabilities to be more competitive in the face of ongoing pressures.
(Randy Kerns is Senior Strategist at Evaluator Group, an IT analyst firm).
Although CommVault had a strong quarter to finish 2012, CEO Bob Hammer spent nearly as much time on the backup vendor’s earnings call this week talking about what is coming rather than what happened.
Without going into too much technical detail, Hammer previewed Simpana 10 ahead of its Feb. 25 rollout. He said the next version of the Simpana backup and management software will include current core technologies such as depuplication and snapshots while adding features to help attract customers in new markets.
Hammer said Simpana 10 will include new features for protecting data on mobile devices and analytics – areas that have not been well-served by traditional backup products — and increased capabilities for storing and managing data in private and public clouds.
Some of the features will include next generation archiving, improved integration with shared services, and more comprehensive reporting and monitoring.
“We’re doing all of this in a way nobody else has done it before,” Hammer said in an interview after the earnings call. “We have massive scale and automation and multi-tenancy, and only make one copy for backup and archive.”
Hammer said the changes are needed to reflect the way customers now protect and manage data, which is created much faster and has more regulatory requirements tied to it than ever before.
“Data management technology has to change,” he said. “With mobile data, customers are having problems protecting data because the technologies they’re using today aren’t adequate to back up in an adequate timeframe. They don’t have time to back it up, and when they do, they have difficulty finding it.”
Part of the scalability for Simpana 10 will be inclusion of an object-based repository, Hammer said. He said Simpana will handle metadata separately, the way object-based storage systems do. “And it’s open, so a third party can write to it,” Hammer said. “You can crawl the network and get data in in a different way than we’ve done in the past. You can search the metadata before you search objects, then do a deeper dive on the content itself.”
CommVault is doing fine with Simpana 9. Its $128 million revenue last quarter was up 24% over the previous year and 8% over the prior quarter. Hammer said he expects to hit $1 billion in annual revenue in a few years, with Simpana 10 the catalyst.
Fusion-io has greatly benefitted from Facebook and Apple for more than two years, with those companies buying tens of millions of dollars’ worth of Fusion-IO PCIe flash cards almost every quarter. Now Fusion-io is seeing the downside of having two customers make up most of its business.
Fusion-io dropped its revenue forecast for this quarter to $80 million from a previous estimate of around $137 million because Apple and Facebook won’t make any meaningful purchases this quarter or next. Fusion-io expects to lose from $10 million to $15 million this quarter after making $13.9 million last quarter. Its stock price fell 22 percent Wednesday after it revealed the forecast drop, and fell another 13 percent today.
Last quarter, Facebook bought $41 million of Fusion-io products and Apple bought $19.3 million, making up a little more than half of the Fusion-io’s $120.6 million in total revenue. In the previous quarter, Facebook and Apple bought $33 million worth of Fusion-io products apiece.
Now the spigot has been turned off, at least temporarily. Fusion-io CEO David Flynn spent a lot of his earnings call Wednesday assuring analysts that Apple and Facebook will return to their status of prolific flash buyers after the six-month lull.
“This is really about the timing of when they put in new infrastructure, not whether or not Fusion-io is a key part of that infrastructure,” Flynn said.
He said Fusion-io has proven its value to those companies, adding “when the pull the trigger for deploying is not in our control.”
Fusion-io could have other problems in the coming months. EMC is pledging more PCIe-based flash products following its VFCache caching software launched in 2012. And Seagate this week invested $40 million and signed a reseller deal with Fusion-io rival Virident Systems. Seagate will sell Virident PCIe flash products through its OEM partners and channel partners.
Flynn shrugged off the Seagate-Virident news, saying “the business is not driven at the OEMs. The business is driven at the end-user customer. So having companies that build components, really they won’t be at the point where the competition is actually happening, and that’s with the end users.”
On a more optimistic note, Flynn said the ioScale cards Fusion-io launched this month are off to a good start. The high-density low-cost ioScale cards scale to 3.2 TB and are aimed more at cloud providers and telcos than enterprise’s mission-critical apps. Flynn said ioScale products were involved with five of the vendor’s 10 $1 million-plus orders last quarter.
Hard-drive giant Seagate is carrying out its flash strategy one piece at a time. This week it added PCIe flash to its portfolio by making a $40 million investment and signing a reseller deal with Virident Systems.
Seagate will sell Virident’s Flash Max II PCIe solid-state cards and will jointly develop new products with Virident. Seagate is Virident’s fourth strategic investor, joining Intel, Cisco and an unidentified large storage vendor that industry sources say is EMC.
Seagate’s flash strategy is to partner rather than develop products on its own. It sells Pulsar solid-state drives (SSDs) through a co-development deal with Samsung and it has invested in flash controller startup DensBits.
“We’ll have a stack of solid state solutions that span from the client all the way through to the flash server high-application data demands of the data center,” Seagate chief sales and marketing officers Albert “Rocky” Pimental said Monday on Seagate’s earnings call with analysts.
Virident launched its Flash Max II single-level cell and multi-level cell cards in August 2012, ranging in capacities from 550 GB to 2.2 TB. Virident’s main competitor is PCIe flash leader Fusion-io.
In a report on the Seagate-Virident deal, Objective Analysis semiconductor analyst Jim Handy wrote that he expects more than two million PCIe SSDs to ship by 2016. He expects PCIe SSDs to outsell SAS-based SSDs.
Seagate has been criticized for moving slowly in the SSD market, but Handy attributes that to the vendor’s exhaustive testing of flash products before releasing them. “With this move Seagate has moved from being a non-participant in the PCIe SSD space to an enviable position with one of the highest-performing products available,” Handy wrote.
Virident now has $116 million in funding from venture capitalists and strategic investors. It closed a $26 million round in September when Mike Gustafson replaced founder Kumar Ganapathy as CEO.
Actifio bills itself as the “radically simple data copy storage company,” but apparently it’s pricing and packaging models weren’t radically simple enough. So Actifio 5.1 released today has new packaging and pricing, including a turnkey solution combining storage and software, a gateway version without storage, and a utility pricing model.
Actifio claims it made 62 sales in the fourth quarter of 2012 for an average price of more than $200,000 with its appliances that change the data protection process. But the pricing could confuse customers used to more traditional backup.
“Customers told us, ‘It can be difficult to comprehend exactly what I’m buying,’” said Andrew Gilman, Actifio’s senior director of global marketing. “So we spent time crafting our packaging. We have pre-packaged configurations that make it easier to understand what’s showing up on the data center floor, and how it can be deployed quickly.”
The Actifio 100T turnkey system comes in eight standard configurations, ranging from 7 TB of protected data for remote offices to 100 TB for enterprises. Customers can stack 100 TB configurations up to 2 PB.
Actifio Gateway lets managed service providers use their existing hardware to build clouds and services based on Actifio’s copy data technology. Its software provides multitenancy and is designed to work on a mix of virtual, physical servers and storage devices.
The utility pricing model lets customers pay as their data volumes grow. It includes software, storage, maintenance and installation, starting at around $4,000 per TB per year for 30 TB and dropping to around $600 per TB in larger volumes.
Gilman said Actifio began piloting the utility model in late 2012, and it is now available on all models with more than 30 TB. Customers can sign up for three-year or five-year subscriptions.
Gilman said Actifiio will let its channel and service provider partners decide whether to use the utility model for their customers.
Actifio also added features in 5.1. They include bare metal restore to enable rebuilds of Windows and Linux hardware, the ability to non-disruptivet failover testing, and the ability to failback and SyncBack after a disaster by only moving changed application data from the DR site to the primary site.
EqualLogic founder Paula Long has a new storage startup, and she says its goal is to do for data scientists what EquaLogic did for storage administrators.
Long was one of three founders of iSCSI storage pioneer EqualLogic, which proved Ethernet can play a major role in network storage. She teamed with former EqualLogic marketing VP John Joseph a year ago to start DataGravity, which today pulled in $30 million in Series B funding to give it a total of $42 million with no products due for at least a year.
DataGravity CEO Long and president Joseph won’t give too much detail on what the company is up to, except to say it’s about gaining business intelligence from data in an easier way than is possible today.
“At EqualLogic our mission was to put an IT administrator in every storage array so an IT generalist could become a storage expert quickly and not spend a lot of time managing storage,” Long said. “We also went for the all-inclusive approach instead of dim sum although everybody thought we were nuts. We took a complicated solution and made it simple.
“Now we’re taking storage experts and marrying them with analytics and database experts, and deriving what happens when you put all those people together. We’re taking EqualLogic’s core principles of simplicity and everything packaged together, and moving that into delivering data intelligence in a simple way.”
The dim sum approach she referred to is the way the major storage vendors sold all of their software features as separate licenses. EqualLogic included all of that in the price of the array, which fit well with its strategy of offering Ethernet storage that was easier to manage than Fibre Channel and didn’t require expensive FC networking devices. EqualLogic was poised to go public when Dell acquired it for $1.4 billion in 2008.
“The last time we looked at iSCSI, SAS, SATA, and all the key enabling tools,” Joseph said. “In this company, we’re looking at a different level of the problem. The focus is now on data and the value that data brings to someone’s industry.”
Long said she expects DataGravity to start shipping product in 2014. She did say it would sell devices and not just software. When asked if the systems would be similar to big data analytics products such as Oracle Exadata, EMC Greenplum, and IBM Netezza, she said “I don’t think there’s anybody similar to us. I’ve seen slices of what we’re doing, but not the same solution.”
Venture capitalist firm Andreessen Horowitz led today’s round, with previous investors Charles River Ventures and General Catalyst Partners also participating. Andreesen Horowitz partner Peter Levine joins DataGravity’s board.
Long said DataGravity went for a B round early to get Andreesen Horowitz and Levine on board. “We’re frugal New Englanders,” she said. “We weren’t planning on raising money until the fourth quarter of this year. But Peter approached us. We thought having him join us early so he can help shape the company would be a great help.”
Levine is the former CEO of XenSource and executive vice president of Veritas Software long before Symantec acquired it. He’s also a current board member of startup Actifio. He said a startup’s leadership plays a key role in his decisions to invest.
“The reason we invested in this company is because of Paula and John,” he said. “No one in storage is more strategic or better able to articulate the storage market than Paula. The concept of turning data into information is quite profound.”
When asked what challenges DataGravity faces, Levine said “Getting a product out is always a challenge. And any time you build a new category, there’s the education and evangelism involved with going after a new market.”
Taking a step back and looking at the new or updated storage systems coming out reinforces trends I see for storage hardware in the industry. By storage hardware, I’m referring to physical system components such as the controller cards, processors and data flow electronics.
The majority of the underlying hardware used in storage comes from standard components. Many of the available storage systems use Intel-based server designs. The value to the vendor from the sale of a storage system comes primarily from the embedded software running on that hardware and associated server software used in conjunction with that storage system.
The embedded software features and capabilities that differentiate one storage system from another are the most focused on product characteristics. For most storage vendors, the software-based features are categorized and put into suites of functionality that have additional charges over and above the basic system costs.
Storage systems that cannot continue to add unique value gravitate to a price based on the hardware costs with some added margins. Characterized as a “race to the bottom,” the undifferentiated systems are generally referred to as commodity storage.
The view that storage hardware is generally moving to commodity may be a natural evolution but is personally disappointing. The manufacture of the common platform for storage systems has moved to lowest cost regions of the world. The equitability of low wages is a different topic, but the commoditization does have a direct relationship with larger scale production and magnifies cost reduction value.
There is a parallel that comes to mind, which is the U.S. steel industry. The technology and production processes for the steel industry were developed and refined in the U.S. over time. But production has moved to lower-cost regions in the world based on price pressures. With the movement of a manufacturing industry, the investment to build production facilities is used to modernize the production.
The surface level similarity between the storage hardware system base and the steel industry has several factors: production moved to follow lowest cost manufacturing, underlying technology was not changing, and investments in the industry were to make manufacturing more efficient.
There still are custom hardware based storage systems in the industry, but they are a minority and the investments in new ideas and hardware technologies seem limited. The competitive price pressures and the continued investments required for innovation make it more challenging to sustain a company with novel technology.
(Randy Kerns is Senior Strategist at Evaluator Group, an IT analyst firm).