Can a new CEO pull Violin Memory out of its current slump?
When you hit rock bottom the only way is up. So the board hopes...
Analysis Is flash array supplier Violin Memory a lost cause needing to look for a white knight acquisition or can it be repaired, see off the competition and return to profitability?
Violin Memory is in a dreadful situation, making heavy and growing losses - $34m and counting - that dwarf its slower-growing revenues, and with impatient investors frustrated by a post-IPO share price some 60 per cent lower than the $9 IPO price set in September last year. The shares are currently trading at $3.80.
The board sacked CEO Don Basile in December, made chairman Howard Bain the interim CEO and announced a search for a permanent CEO. Bain set about stabilising the company and, hopefully, stemming its losses.
It is being assailed by an activist investor the Clinton Group, which is telling the board to sell the company. However the Clinton people don't seem to have built up a sufficiently large enough holding to command attention from the board.
Basile's departure was only the latest in a stream of exec changes and departures and was itself followed by two more resignations; Matt Barletta, the VP for marketing, has left the company. He had been regarded as one of Don Basile's people, along with chief operating officer Dixon Doll (who bailed out last week), and has, so The Register's storage desk understands, left on very good terms. That makes four VP-level and above executives who have left the company in less than a month:
- Software CTO Jonathan Goldick on 12 December
- CEO Don Basile on 16 December
- COO Dixon Doll on 2 January
- VP marketing Matt Barletta also in early January
Violin has had rapid EMEA boss changes too, having gone through three EMEA heads in less than a year, finishing up with Mick Bradley's appointment in November 2013. The European channel had to get used to working with yet another EMEA MD as Basile hired and fired EMEA bosses searching for the magic guy who would get Violin's kit moving in the territory.
How did Violin, which is generally reckoned to lead the all-flash array market, get in this hole and how might it climb out?
The company has a line of 3000 and 6000/6200 all-flash arrays. These use its own, non SSD-based technology. It has also entered the PCIe flash card and SAN acceleration markets yet has relatively weak software that it can call its own.
To use a cliche, one of its greatest strengths has been the drive and forcefulness of ex-CEO Don Basile, while one of its greatest weaknesses has been .... the drive and forcefulness of ex-CEO Don Basile. He excited strong opinions in other people.
We have been told that some wouldn't cross the street with a bucket of water if he was on fire. Another source added that under his leadership, Violin's management of expectations with Wall Street and the investment community was poor; and that its sales management was lacking – at least until Dixon Doll was promoted to COO from head of the sales organisation.
Let's have a look at some of the issues Violin has had.
Getting on the wrong side of HP
Violin used to supply flash arrays through HP so that HP could compete with Oracle and others. Unfortunately it was partnering Business Critical Systems – old HP, so to speak – and when HP bought 3PAR a bit of a war started up inside the company.
Should HP OEM and partner for flash arrays or develop its own? Itanium-based systems were dying on their feet, and 3PAR boss David Scott pushed for the in-house development. Violin's HP contacts and its relationship weren't strong enough to either destroy this 3PAR-led strategy or somehow work alongside it. So HP pressed the eject button and Violin lost its biggest and probably its best reseller.
Violin's board sanctioned and approved Don Basile's strategy and activities until late last year when the post-IPO debacle and losses outgrowing revenues forced its hand, meaning Basile had to go.
The board now consists of :
- Howard Bain III - Chairman and interim CEO following Don Basile's sacking as CEO in December.
- David Walrod, a venture partner at Bridgescale Partners.
- Larry Lang - described as: "currently the President and CEO of QuorumLabs Inc. QuorumLabs develops solutions that make world-class business continuity easy and affordable for enterprises of any size. ... Previously, Mr. Lang served as vice president and general manager of the mobile service provider business unit at Cisco."
- Jeff Newman - listed as "a longtime Silicon Valley executive, Jeff brings his vision, leadership qualities, and depth of strong alliances and relationships with an industry-wide global network. ... Experienced Board member to many client companies, Jeff has helped shape optimal corporate roadmap and growth strategies for 50+ companies."
- Don Basile - the former CEO is still listed as a board member on Violin's website.
- Cheemin Bo-Linn was elected to the board as a director on 20 December 2013. She is the President of Peritus Partners, an “international consulting group focused on leading companies to the next level of growth and increasing business valuation.” She was elected to ensure Violin has enough independent directors and committee chairs to satisfy NYSE rules. When Bain became CEO on 16 December, he lost his independent director status.
With the exception of Cheemin Bo-Linn this is the board that was in place before and during the IPO and subsequent period. It is a Basile-era board. In fact, any incoming CEO may well request and even require board-level changes.
It's noticeable that neither of Violin Memory's founders are on the board, meaning Jon Bennett, the CTO for hardware and technology, and Donpaul Stephens, Violin's president. Stephens is not included in Violin's management team listing either.
This board had better be prepared for changes; investors believe it has screwed up, so some long-term members are likely to go.
Product problem areas
Violin has built itself a good flash array business and forced itself to the top of the all-flash array tree but now questions are being asked about product strength, particularly on the software side, and core product dilution, as mainstream competitors position to assault Violin and try to eject it from their customer bases - EMC, Dell, Cisco, HDS, HP, IBM and NetApp are the main ones. Startups Solidfire and Pure Storage look to have strong product offerings and better marketing. The low-hanging all-flash array fruit days are over.
Violin has perhaps spread itself over too many product areas, with PCIe flash and SAN acceleration cache alongside its core flash arrays, and paid too little attention to developing its own software resources.
Was Violin's development of a PCIe product line a mistake? It is an already crowded market, with Fusion-io, IBM, Intel, LSI - now part of Avago, Micron, HGST with Virident, and several others. If you don't have server manufacturer or storage array supplier OEM deals then your product is a server retrofit, one for white box manufacturers (like Quanta and Supermicro), independent system vendors and savvy customers. The product has to stand out. It has to perform excellently, be managed simply, work with hypervisors, cache or tier hot data through good software, and work with back-end arrays.
It's not clear yet whether Violin has developed such a product or merely diverted resources away from its mainstream 3000 and 6000 series flash arrays with a low priced but ho-hum Velocity PCIe flash card line.
The company is also developing a PCIe card product for Toshiba, its flash foundry chip supplier and that is additional engineering development effort.
Converged flash array and servers
Violin has quite a history of running system software inside its arrays. In August 2012 it signed a deal with VMware to run vSphere inside them. At the time, Violin product VP Narayan Venkat, said: "This is one of the many steps we've taken to apply the benefits of flash to virtualised environments, and create specialised application acceleration appliances like Virtual Database-in-a-box, Private Cloud-in-a-box and VDI-in-a-box."
In October last year Violin announced it could run Windows Server 2012 R2 in its 6000 and 6200 arrays. We wrote then that it was "giving the server OS access to up to 64TB of NAND with the ability to run all Windows enterprise applications in solid state storage."
So here we had a converged server and flash storage array that could give its applications access to data with no network access latency, ensuring they performed much better than systems with servers fetching data across a network link from flash or disk-based arrays.
Curiously, earlier that year Violin had partnered with Atlantis to run with its ILIO VDI system. This involved the ILIO software running in a separate server with a network link to the flash system. The ILIO software did not run in the Violin array, thus putting Violin at a potential disadvantage versus hybrid array makers Nimble Storage, Tegile and Tintri, who could use cheap disk for bulk storage and flash for the fast VDI work. Violin, with its all-flash array, was restricted to the large scale VDI market which has yet, we understand, to take off.
Venkat's VDI-in-a-box idea had not come to pass with the Atlantis deal.
Violin isn't positioning its products as converged server/storage systems like those from Nutanix or Simplivity. It could, if it wished, add more server and DRAM resources, and do so but this would represent another product line and marketing effort. There is a decision to make here; is the company making all-flash arrays or converged flash storage/server systems?
With a PCIe product line to look after as well we might well think that Violin's strategy might better be formulated to look after its core flash array products and not dilute its resources with side efforts on both PCIe and converged systems, and SAN caching.
After buying GridIron in January 2013, Violin developed its Force 2510 appliance, announcing it in September. This is a flash cache sitting in front of a (disk-based) SAN and accelerating its data access with its hardware and Maestro software. It also transparently copies data from the SAN to a 6000 series Violin flash array.
This is another product line development, like PCIe flash, and server/storage convergence, that has yet to come to fruition. We'll see, hopefully, how it has done in Violin's next quarterly results statement, expected in late February.
Outsourcing software development
Violin, it is generally agreed, lacks its own system software features, such as deduplication to increase storage efficiency, thin provisioning and more. It has a deal with Symantec for resale of that supplier's Storage Foundation software which does have deduplication, compression and thin provisioning, and also automated storage tiering. Symantec has recently made Storage Foundation flash cache-aware to better use server flash caches.
FalconStor is being funded by Violin to produce a new version of its NSS storage virtualisation software that makes better use of flash memory. FalconStor recently dismissed its own CEO and has a relationship with a venture capitalist skilled in restructuring. It looks like a weak software development partner.
Symantec too has a relatively new CEO, Steve Bennett having replaced Enrique Salem in 2012, and has its own struggle to return to a good rate of profitable growth.
Why hasn't Violin concentrated its resources on developing its own software? Both Symantec and Falconstor can sell their software elsewhere, diluting Violin's software uniqueness. Perhaps this software outsourcing was part of the reason for Goldick's departure.
We asked ESG founder and senior analyst Steve Duplessie whether Violin should seek a sale or whether it should get a new CEO as soon as possible and refresh/restructure its product lines?
His answer was this: "All good questions. It put itself between a rock and a hard place going public prematurely in my opinion. Now it has to resurrect itself in the spotlight, which is brutally tough to do. But it can be done; the market is still red hot. They need to put grown-up processes in place, get the product rock solid, and find a way to add some software value. If they can do that, they have just as good a shot as anyone else does. A new CEO without the outlandish arrogance should be a breath of fresh air for the employees. So who knows? But I'm not ready to write their potential off just yet."
Steve O'Donnell, until recently a member of Violin's advisory board, said: "Violin is the leading all flash storage business with a rock solid and highly performant set of products. The market seems to be comparing them to commodity players like Fusion-io when their biggest competitor is clearly EMC with their premature product XtremIO.
"EMC was forced to deliver XtremIO early to have a chance of answering Violin in the market. Why? Well, because EMC were being challenged by Violin at the top end, in all of the sales where EMC make most profit. Lots of EMC folks are really upset that a product is on the market that hasn't been properly bedded down, risking the EMC brand," he said.
He opined: "I find it incomprehensible that EMC ship a product that keeps critical metadata in RAM that needs battery backup to survive an outage. This is a disaster waiting to happen."
O'Donnell added: "Pure Storage is challenging the mid-range and Violin the top end of EMCs products and the opportunity for massive market disruption is clear."
There could be a new CEO announcement at Violin as early as this week. A consensus view is that the company can be restored to an even keel.
Violin is still, overall, a great business, with an effective sales team, said another source, but it needs a new leader - fast - and a renewed focus on getting its message out. No one should write it off. ®
Sponsored: Hyper-scale data management