Datrium has added an “Insane Mode” feature that the startup claims can double host storage performance on the fly.
Datrium took the Insane Mode tag from Tesla’s rapid acceleration technique for its Model S cars. The concept is similar – the goal of Datrium’s Insane Mode is to let customers instantly speed the storage systems performance by increasing available CPU resources.
Here’s how Datrium DVX works:
The system has two parts. The Diesl (Distributed Execution Shared Logs) file system runs on the host and NetShelf storage array. Diesl allows customers to manage storage like virtual machines through VMware vSphere, and handles features such as compression, deduplication and instant zero-copy clones. It uses up to 8 TB of NVMe flash in the server to boost performance. Customers can use any commercial flash in the host.
NetShelf is the second part, and serves as durable storage. NetShelf has dual controllers – each with NVRAM – for active/passive failover and hard disk drives for capacity. It exports no blocks or files and has no LUNs or volumes to manage. Diesl provides host-based data services. Writes are mirrored before control is returned to a virtual machine. A failed host does not impact fault tolerance or cause data loss. A NetShelf includes 29 TB of data after RAID 6 overhead.
Insane Mode an alternative to QoS
Insane Mode is the third method for improving DVX performance. The others are to add flash to the host or to vMotion a workload to another host with more available flash.
Insane Mode, now part of Diesl Hyperdriver Software, reserves 40% of host cores for DVX instead of the default 20%. Datrium recommends using Insane Mode only when a host averages less than 40% CPU utilization for VMs because ESX utilization should be kept below 80%. Customers can stay in Insane Mode indefinitely or use it for specific tasks like running batches before returning to 20% utilization. Customers can go from Fast Mode (the default) to Insane Mode with one mouse click from the DVX dashboard.
Datrium marketing VP Craig Nunes said doubling the CPU resources can improve workload performance from 1.5x to 3x “literally instantly.”
He said the feature is an alternative to quality of service on traditional arrays. “A storage array has a set amount of performance in its controller. When you hit the boundaries, it’s upgrade time,” he said. “Quality of service on the array is about how you provision precious resources in a storage array. The approach we take is, service levels are more about giving more resources for performance instead of limiting precious resources.”
Datrium CTO and founder Hugo Patterson said the extra CPU is available because customers rarely go above 50% utilization. “If you’re not doing I/O, then CPUs are available for other things,” he said. “This basically reserves enough so when you need to use it, it’s there.”
Patterson said he expects Insane Mode to be used for mainly for test/dev, virtual desktop infrastructure and server consolidation.
Datrium bucks all-flash, hyper-converged
“Flash belongs on the host, and flash has been on the host for a decade,” Patterson said. “But when it is on the host, it’s just a storage device. It’s an island of flash, not an enterprise class storage system. One SSD is not an enterprise class storage system. So how do you get flash on the host and yet make it an integral part of an enterprise storage system? That’s what the DVX is really all about. We’ve moved a lot of the storage compute – just about all of it – into the host. These days it’s a bigger deal to do that with the storage compute getting bigger and bigger.”
Patterson also argues that a DVX costs less and uses both the server and storage better than a typical hyper-converged system.
“Each server is independent [with DVX] and gives an easier way to manage storage performance,” he said. “Communication is from servers to NetShelf and servers don’t talk among themselves, except in the case of vMotion. We don’t have one host accessing data that’s on another server. When you have a cluster of [hyper-converged] servers, they’re all co-dependent and they can heavily influence the performance of each other. When performance isn’t what you hoped for, it’s hard to do anything about it. With DVX, each server has its own local cache and uses its own CPU.”
Brocade’s earnings for last quarter show a decline in storage sales consistent with what large array vendors have reported.
The Fibre Channel switch vendor’s forecast for this quarter shows that decline is likely to continue.
As the leading FC switch vendor, Brocade’s revenue is tied into the sale of SAN arrays. With the industry in transition to cloud storage, server-based storage systems such as hyper-convergence and new all-flash arrays cutting into adoption of traditional FC storage arrays, storage vendors have seen revenue decline over the past year. Brocade CEO Lloyd Carney said on his Thursday earnings call that he expects FC storage to bounce back by the end of the year but not before it declines more than previously expected this quarter.
Carney said the steep decline in Fibre Channel revenue hit directors and switches. “SAN revenue was weaker than we expected,” he said. “The weakness was broadly spread across our partner base and appears to indicate lower overall storage spending.”
Brocade’s reported revenue of $523 million decreased four percent from last year and nine percent from last quarter. FC SAN revenue of $297 million fell five percent year-over-year. IP (Ethernet) networking revenue – roughly one-quarter of Brocade’s business – fell nine percent year-over-year to $132 million. But Brocade forecast an increase of seven percent to 14% in Ethernet sales this quarter due to a swift uptick federal and enterprise spending. Meanwhile, it expects FC sales to drop from 3.5% to seven percent over last year to $510 million to $530 million. Carney said he does expect a rise in FC spending next quarter and predicted revenue declines for the full year will be around two percent.
“There are a lot of really big deals in the pipeline,” Carney said of his optimism for next quarter.
Brocade is counting on 32 Gbps switching products and flash storage to give FC a long-term jolt but that hasn’t happened yet. Flash arrays sales are picking up but not the switching that goes with them, and 32-gig (Gen 6) FC won’t become mainstream before 2017.
Brocade is shipping its first 32-gig FC switch, the G620, with 32-gig directors and embedded switches expected in August. But bandwidth upgrades usually occur slowly, and end-to-end 32-gig support will also require upgrades from storage array vendors.
“We see good interest in [Gen 6], products are being well-accepted, and we’re pretty excited about it,” Carney said.
Flash arrays are becoming popular now, with most storage vendors reporting large increase in all-flash systems even as their overall storage sales drop. But Brocade executives said customers are still planning their upgrades and how they will affect their switching.
“It’s causing some of these customers to have to take a look at their architecture, see how these build-outs happen going forward, and there is some bit of a pause [in buying],” Carney said.
“Although all-flash has been here for a couple years and grown pretty fast, it’s really been a small part of the overall storage marketplace. Now, that’s changing because all the large players have all-flash arrays, and they’re growing very fast. They’re putting a lot of resources behind it. For us, the good news is most of them connect using Fiber Channel. Now, they’re more efficient, so they need less storage. But storage is still growing. So it kind of works out.”
Carney said Brocade also hopes to receive a sales jolt from hyper-convergence. Hyper-converged systems are seen as harmful to FC because they often replace SANs. But Carney said he expects that to change as hyper-converged rack-scale products such as EMC’s VxRack and the Hitachi Hyper Scale-Out Platform pick up steam.
“If you are buying individual hyper-converged platforms, you don’t need us as part of the solution, the storage is already in there,” he said. “If you want to connect two or three hyper-converged boxes together, the best way to do that right now is Fibre Channel.”
The product name Unity is at the center of a trademark dispute in the data storage industry between Nexsan and EMC.
Nexsan Technologies filed a complaint against EMC on May 6 in the U.S. District Court in Massachusetts claiming it has priority to the Unity trademark and is not infringing upon any trademark rights that EMC may assert.
The lawsuit came less than two weeks after both companies orchestrated major product launches around Unity product lines. Nexsan, a wholly owned subsidiary of Imation Corp., issued a press release on April 26 for its new Unity storage line, combining NAS and SAN storage and enterprise private-cloud file system synchronization.
On May 2, EMC launched its all-flash Unity mid-range array at EMC World, providing unified block and file storage. EMC said it would also offer a hybrid version of Unity with hard-disk drives.
In its May 6 lawsuit, Nexsan noted that it filed trademark applications for the terms “Unity” and “Nexsan Unity” on March 22, 2016, with the U.S. Patent and Trademark Office (USPTO). Nexsan said EMC did not file trademark applications for the “Unity” and “EMC Unity” marks until more than a month later, on April 29.
In an email to Nexsan dated April 29, EMC Intellectual Property Counsel John Hurley noted that Nexsan had filed two trademark applications, but he claimed Nexsan had filed “no amendment to allege use” in support of either application.
Hurley wrote that EMC first used the mark “Unity” for an extension of its VNX product line more than a year ago in customer presentations. He claimed a presentation to a customer and large reseller on March 19, 2015, and at least three more customer presentations in that month alone.
In the April 29 email, Hurley asked for Nexsan’s immediate agreement to:
–“Expressly abandon” its trademark applications with the USPTO for “Nexsan Unity” and “Unity.”
–“Not use any UNITY mark for goods or services related to EMC’s UNITY goods and services.”
–By May 6, send a copy to EMC of its submission to the USPTO to withdraw or abandon its two Unity applications.
“Based on EMC’s senior use of its UNITY marks, we believe Nexsan does not have the right to use UNITY as a mark on similar goods and services,” Hurley asserted in the email to Nexsan.
Hurley added: “EMC reserves the right to commence legal action without further notice in the event those applications proceed toward registration and/or Nexsan initiates use of any UNITY mark in connection with similar goods or services.”
Noting that EMC “threatened suit,” Nexsan filed its lawsuit against EMC on May 6, claiming that EMC had not used the EMC marks “in commerce for services” since 2015, as Hurley claimed.
“Priority is based on who uses or files the mark first, so you can have rights based on filing first or based on using first. That’s literally what this lawsuit is about,” determining if Nexsan filed its application before EMC used its mark in commerce, under the requirements of U.S. trademark law, said Lisa Tittemore, an attorney at Boston-based Sunstein Kann Murphy & Timbers LLP, the firm representing Nexsan.
Nexsan CEO Bob Fernander said the company has no intention of dropping the name Unity. He said he had no knowledge of any 2015 EMC customer presentations in which the term Unity may have been used.
“You’ve got to do it publicly,” Fernander said. “And the definition of public is kind of common sense, we think. And your commerce becomes another next step in the process of making it publicly known that you’re using a mark. So we’re scratching our heads on that one.”
Fernander said he first learned of EMC’s Unity product after the product launched. “We got a call from a reseller of ours who said, ‘Hey, do you guys realize that EMC just launched a product that’s the same name as yours?’ ” he said.
Nexsan conducted a poll via Survey Monkey among employees to select the name of its new Unity product, according to Fernander. He said employees had to choose among a list of potentially suitable names.
Would Nexsan consider a settlement? “Anything’s possible,” said Fernander, “but right now, I represent a group of shareholders for Imation and Nexsan, and it’s my job to protect their interests.”
EMC has yet to respond to Nexsan’s lawsuit in court and declined an interview request, saying it does not comment on pending litigation.
Data Protection specialist Datto is giving away its new file sync and share service for a year.
The Datto Drive service for SMBs is hosted on the vendor’s 200 PB private cloud within nine worldwide data centers. The service comes with a global license agreement with open source file sync and share vendor ownCloud.
One million Datto Drive accounts will be offered for free. After a year of service, customers will pay $10 per terabyte per month and a Datto managed service provider partner (MSP) will facilitate the relationship with the customer. Datto sells exclusively through MSPs.
“We are trying to facilitate the health of the (SMB) ecosystem because it is critical to our businesses,” said Brooks Borcherding, Datto’s chief revenue officer. “We see file sync and share as a commodity market, yet it can be expensive for small businesses. Fifteen dollars a seat is quite expensive.
“We know we are going in to a commoditzed space, so it’s very competitive. We thought we could go into it in a disruptive fashion and make some noise.”
The Datto Drive, which became generally available May 2, is a full software as a service, multi-tenant offering that is managed on ownCloud so customers can connect and ingest files from other FSS vendors such as Dropbox and Box. Customers can access, control and manage files from any device. It supports the iPhone, Android, iPad, Microsoft Windows, Linux and Mac.
Customers can choose between two models, that both include up to 5 TB of storage. The Datto Drive Business Edition. includes year-round support, and replicates between data centers and the cloud nodes for a tertiary-level of redundancy. The Professional edition allows MSPs to customize its logo.
Datto’s core business is backup and disaster recovery as a service. It sells a hybrid backup appliance to SMBs for local backups and replication to the cloud. The Norwalk, Conn.-based company has raised $100 million in total funding with international offices in the U.K, Australia and New Zealand.
Like many storage vendors, Quantum’s revenue took a dip last quarter and for the full year. The backup and scale-out storage specialist predicts it will rally beginning this quarter, however, due to a pickup in video-related storage systems.
Quantum’s total revenue of $120 million for last quarter was down from $147.8 the previous year. It was Quantum’s fourth fiscal quarter, and its revenue for the year of $476 million was down from $553 million the previous year.
Quantum lost $50.2 million last quarter and $74.7million for the year, mainly due to a $55.6 million goodwill impairment charge. Its non-GAAP loss for the year was $2.6 million and it had a non-GAAP profit of $6.5 million for last quarter.
Quantum’s traditional products are in data protection – tape and disk backup. But its scale-out storage is its fastest growing segment and was approximately 27% of its total revenue last year. The Quantum scale-out portfolio is driven by its StorNext file system, which also powers its Artico NAS appliance and Xcellis workflow storage. Quantum CEO Jon Gacek said he expects that product line to rapidly grow to cover use cases such as video surveillance, media and entertainment, life sciences, and sensor-based Internet of Things applications.
Quantum’s $33.1 million in scale-out revenue was a four percent increase year-over-year, and the vendor forecasted higher growth this year. Scale-out revenue of $126.5 million for last year grew 23%year-over-year, and Quantum executives said they expect that to grow to around $175 million to $200 million of its total $500 million revenue for this fiscal year. The expectation is also for a modest increase in disk backup and declines in tape backup for an overall decline in data protection revenue.
CFO Fuad Ahmad said Quantum added around 110 new scale-out customers last quarter and approximately 460 for the year.
Scale-out revenue increased despite a $7 million drop in deals of $1 million or more.
“We delivered $120 million … with the scale-out not having its big growth quarter,” Gacek said Tuesday on the earnings call. “To do a sports analogy, we won the game, we covered the spread and our best player didn’t really play well. The best player in this case is scale-out and I just think the scale-out opportunity will be there for us. We’re super well teed up.”
Quantum’s tape automation revenue of $44.7 million last quarter was down from $61.4 million the previous year, and disk backup revenue of $18.1 million in the quarter fell $7.1 million from the previous year.
In an interview after the call, Gacek said the opportunities in scale-out storage are greater than in traditional storage markets. “The traditional storage side of the company is still feeling pressure, but it’s the scale-out storage I see growing – there’s a lot of video where performance and access are important,” he said.
Gacek said Quantum has pulled back its investment in DXi disk backup, a platform that it has aggressively upgraded over the years. “There’s modest growth there,” he said. “We don’t go out of our way to spend a lot of money on DXi. We will upgrade servers and storage and stay current with application support, as opposed to having the fastest deduplication engine.”
He added that backup and scale-out products are often combined into large deals, where customers start out looking for backup and end up adding archiving for unstructured data.
“A lot of challenges of unstructured data can be solved with archiving,” he said. “Why pay a bunch of backup licenses when you can archive it?”
Gridstore chairman Nairman Teymourian took over as CEO today, with plans to install a utility services model for delivering hyper-convergence.
Teymourian has been Gridstore’s chairman since October. He replaces George Symons, who left soon after Teymorian joined and Gridstore closed a $19 million funding round in January. Teymourian had been general manager of Hewlett-Packard’s converged systems business for 20 months before becoming Gridstore’s chairman. He has also been CEO of Gale Technologies, Caspian Consulting, Prism Health Software and Software Reliability. All those companies were acquired by larger organizations.
But Teymourian said he plans to grow Gridstore by expanding its Microsoft Windows-based hyper-converged business.
“I believe the infrastructure market is going to be disrupted by hyper-convergence,” Teymourian said. “And we can disrupt the current hyper-covnerged players. I think the real value of hyper-convergence has yet to be delivered.”
The way to do that, he says, is through a pay-as-you-go model.
“Gridstore today has a very good product,” Teymourian said. “It’s a platform that is fully integrated with Microsoft. The value is it is simple, fast and inexpensive. It comes out of the box and delivers value fast. Our integration with Windows is unique. Plug us into a wall and we come up as a C drive in Windows. There’s no learning curve. And we’re all flash.
“But I think the world is moving towards a model where compute, network, storage and all physical hardware is going to be commoditized. Customers don’t want to buy hypervisors. What they want is that when the box arrives, they can pay for it when they use it.”
Teymourian said Gridstore will deliver hyper-convergence that costs about the same as storing data to Amazon Web Services (AWS). Gridstore will ship customers an appliance at no charge until the customer buys the Gridstore service that offers choices of hypervisor along with application, management and deployment options. “They can build a workload, deploy a workload and then with a single push of a button, they can move that workload into the public cloud,” he said. “Our plan is to support as many public clouds as possible. In a quarter or two, you’ll see a very different hyper-converged platform in the market.”
Gridstore will continue to support Windows and Hyper-V, but Teymourian said it will add support for KVM hypervisors and containers. Not VMware, though. “VMware has fundamental issues with its licensing,” he said. “People are getting tired of paying a VM tax. They want their hypervisors to be commoditized. Hype-V is free.”
Teymourian said customers will receive a Gridstore appliance and plug it into the network, but will only pay for when they use it with the software stack. He said Gridstore’s control plane can run in the cloud and customers can choose management options. They will also be able to run Gridstore as a “private cloud in a box that looks like the AWS model. Or it can be a managed service – the customer can say ‘I want to dial-up to the box wherever it is and pay for it.’”
He said the customer would not even know what hypervisor is running. Gridstore will also support bare metal hyper-converged deployments under the services option. Pricing would be set according to workloads, CPU cycles, number of VMs supported and other metrics.
LAS VEGAS — EMC dropped a surprise at EMC World today by previewing an all-flash version of its Isilon clustered NAS system.
C.J. Desai, president of EMC’s Emerging Technologies Division, said “Project Nitro,” has been in the works for 18 months and is scheduled to go into beta in the second half of 2016 and begin shipping GA next year. Desai said the system will scale to 400 nodes, 100 PB of capacity and 1.5 TB per second of throughput. He described it as a “super super dense” system packed with 15 TB solid state drives.
EMC currently has four all-flash storage systems with the launch of the midrange Unity array this week. Unity joins VMAX All Flash, XtremIO and the DSSD D5 massive shared storage system in EMC’s portfolio. Isilon will be the last of the vendor’s primary storage systems to go all-flash.
Nitro looms as a competitor to Pure Storage’s FlashBlade array for unstructured data, which was announced in March and is due to ship later this year.
Data protection vendor Commvault Systems generated $159.6 million in revenues in the fourth quarter, hitting the second consecutive quarter of sequential growth after three quarters of declining revenues.
The company beat Wall Street analysts’ expectations by $2.47 million, with total revenues up by six percent year-over-year and two percent sequentially.
Brian Carolan, Commvault’s chief financial officer, said total revenues for the full fiscal year were about $595 million, which were a decrease of two percent compared to the 2015 fiscal year. Total revenues for fiscal 2016 were up three percent year-over-year. Software revenues for the fourth quarter 2016 were $73.3 million, up three percent sequentially and five percent year-over-year.
Revenue from enterprise deals, defined by Commvault as deals over $100,000 in software revenue, were 58% of total software revenue and the number of enterprise deals increased 13% sequentially.
“Our average enterprise deal size was approximately $272,000 during the current quarter, which was down 2 percent from approximately $278,000 in the third quarter 2016,” Carolan said.
Commvault’s net income was $16.6 million for the quarter and $42.4 million for the year.
President and CEO Bob Hammer attributed the growth to changes Commvault made during a two-year transition, which brought about license revenue from new distribution partnerships with Microsoft, Amazon Web Services (AWS), Cisco, Nutanix and Pure Storage.
The Commvault V11 Data Protection Platform, which was re-branded from Simpana and introduced last fall, is experiencing strong sales. The company added about 450 new customers this last quarter and revenue made via its deal with Arrow distributor was 38% of the company’s total revenue, a nine percent increase year-over-year and flat sequentially.
“The move to the cloud has become a major factor contributing to our increased business momentum since Commvault solutions can holistically solve a broad range of critical new problems that customers are facing as they migrate to the cloud, manage data in private, hybrid and public clouds and deploy new hyper converged and big data infrastructures,” Hammer said.
The company is winning larger enterprise deals, Hammer said.
“However, our ability to grow is more dependent on just big deals or the steady flow of $500,000 and $1 million-plus deals,” he said. “These deals have quarterly revenue and earnings risk due to their complexity and timing. Even with improved funnels, large closure rates may remain lumpy. We still face critical challenges. We need to expand the market visibility and understanding of the strategic value of Commvault’s V11 Data Platform and broader set of standalone products.”
CommVault Systems’ history is in data protection but it has been adding cloud data management as part of its core strategy. Hammer said cloud and software as a service (SaaS) vendors to not replace the need for the type of data management CommVault is offering.
“For example, replication data protection technologies from cloud and SaaS vendors do not replace the need for a backup copy for critical data needs such as compliance, legal and historical business analytics,” he said. “They do not provide the data and information capability to federate the management of data across silos of cloud, mobile, SaaS and new IT infrastructures.”
Hammer said Commvault will expand its Data Platform over the next six months or so, adding user self-service for managing cloud data, improvements for managing data created in the cloud, federated indexing, and intelligent search of online clouds and managed copies.
LAS VEGAS – Michael Dell today said his company will be called Dell Technologies after he completes his $67 billion acquisition of EMC. The EMC name will live on for the enterprise business, which will be called Dell EMC.
Dell revealed the name during his keynote address at EMC World 2016. He said Dell, EMC and its pieces, including VMware, Pivotal, Virtustream and RSA, and Dell’s recently spun off SecureWorks will all be part of a strategically aligned family of businesses. “As family names go, I’m kind of attached to Dell,” was his reason for calling the company Dell Technologies.
Dell used the keynote to try and convince attendees that EMC will be bigger and better after the acquisition. He said the deal will close “under the original terms and under the original timeframe.” The deal is expected to close by October. The two major hurdles are regulatory approval in China and ratification from EMC’s shareholders.
Dell compared his company’s direction to Hewlett-Packard, which last year split the company into two. “Companies like HP are shrinking their way to success. Wait, you can’t shrink your way to success. That’s not a reality,” he said. “They’re separating their edge from their core.”
Cinder still has the highest adoption rate among OpenStack storage projects, but interest in the nascent Manila file-share service is picking up, according to a 2016 user survey released this month by the OpenStack Foundation.
The survey showed that 57% of 290 deployers who answered the adoption question use OpenStack Cinder block storage in production, and another 26% are testing it. OpenStack Swift object storage was in production use at 32% of the deployments, with another 21% using it in test mode.
The emerging Manila shared file system had production deployments among only 3% of respondents and test usage at 8%. But Manila is generating lots of interest. Only the Magnum containers service (44%) and Designate DNS service (41%) generated more mentions than Manila (38%) among the 290 respondents who rated the OpenStack projects they’re most interested in.
“We’re starting to see a handful of big deployments move past the ‘kicking the tires’ phase and deploy Manila in production. The number of new users trying Manila is also more than I can count,” said Ben Swartzlander, the OpenStack Manila project team lead and a senior software engineer at NetApp.
The OpenStack Foundation’s volunteer survey generated responses from 1,603 community members representing 1,111 unique organizations and 405 user deployments. The IT industry dominated the survey pool, at 68%, following by telecommunications (14%), academic/research (9%), financial (2%) and film/media (2%). The number of users responding to specific questions varied.
The top priority for most respondents is saving money over alternative infrastructure choices, and the majority of deployments are in on-premise private clouds.
More than 75% of 256 respondents use between five and nine OpenStack projects. Looking at OpenStack storage, 20% of 312 respondents said they use the open source software in production for storage/backup/archiving purposes. Another 5% use it for development/quality assurance, and 3% are involved in proofs of concept.
The top Cinder driver in production use among 260 respondents was Ceph RBD followed by the default logical volume manager (LVM) for Linux (16%), NetApp (8%), NFS (5%), GlusterFS (5%), SolidFire (4%) and VMware VMDK (3%).
And there are some sizable Cinder block storage deployments. About 9% of 148 respondents have more than 1 PB, including . The Cinder breakdown for the rest was:
19% – 100 TB to 999 TB
38% – 10 TB to 99 TB
24% – 9 TB or less
Among the OpenStack Swift respondents, the breakdown was as follows:
4% – 1 PB to 99 PB
20% – 100 TB to 999 TB
25% – 10 TB to 99 TB
51% – 9 TB or less
Asked what kinds of data they plan to store on object storage in the next 12 months, the respondents said:
68% – backup/archiving
60% – Docker/container/VM images
58% – application data
32% – big data
3% – other