This week, Amazon announced a new storage offering within Amazon Web Services (AWS) called Glacier. Aptly named, it’s intended to be vast and slow moving, with a cheap price tag to match. At a fraction of the cost of the storage intended for online storage offerings EBS and S3, Glacier will cost you $.01 per GB per month, compared to around $.05- $.13 per GB per month for higher performance offerings depending on capacity stored. Restores from Glacier are costly by design; this is intended for data that you’re not likely to access frequently. If used for the right types of data, this will be a low cost way to park stale data for long periods of time.
Analyzing the cost implications, it would cost you all of $120 to store a TB of data for a year, provided you don’t have to access it during that time. Ten years would cost you $1,200, and 100 years would cost you $12,000. Sure there would be upcharges if and when you access the data, but the value of being able to get back the data you need within a few hours, years after you archived it is tremendous. The data reliability guarantee is 11 9’s— meaning that for each piece of data, Amazon guarantees that it will be there 99.999999999% of the time, included in the base cost to archive it, which is about as close to certainty as you can get in any contract.
I was at Marc Benioff’s subversive non-keynote at Oracle OpenWorld yesterday, and while it was fun to see all the hoopla (employees holding posters of Benioff cast as a dissident, shouting, honking, donuts, cocktails), it was also cool to have the "I was there when" moment as Oracle’s future biggest competitor draws the lines of battle that are likely to shape the enterprise software industry for the next decade. Truth be told, I think that Benioff was a bit too caught up in the fuss and the cloudwash to make me think he’s a mature and credible competitor yet, but he is clearly getting his gumption up.
Benioff pointed a finger at Exadata as a new mainframe, locking customers into proprietary hardware and forcing them to buy over-expensive gear from an industry monolith. He described his own company as “open,” allowing customers to move to any platform or any cloud; "philanthropic," donating $24 million in grants and using their OOW booth as an engine for giving; and "social," leveraging their internal social media engine Chatter to coordinate their rapid mobilization to deliver the non-keynote within 16 hours of being cancelled by Larry. And all of that is cool, but I think he skewed to tech industry buzz rather than focus on the real competitive forces between Oracle and Salesforce.
Today, Nimbus Data Systems announced that eBay has deployed more than 100 TB of their flash storage capacity, putting the spotlight on a key win for the emerging storage vendor. This is interesting on a few levels. First, it’s good to see some life from the emerging storage vendor space, after the pool has been depleted by numerous acquisitions by major storage vendors over the last couple of years. This announcement will get Nimbus more attention and give it some headway in the race to survive and succeed in the roster of next-generation of storage system vendors that includes Arkologic, BridgeStor, Caringo, Cleversafe, Coraid, Nexenta, Promise Technology, RELData, Scale Computing, Scality and Stormagic, as well as the crop of cloud gateway providers, Nasuni, Panzura, StorSimple and TwinStrata (I’m always trying to expand this list, comments or suggestions are always welcome…).
I was recently taken to task in the twittersphere (I post as @reichmanIT) by @Knieriemen, a VP at virtualization and storage reseller Chi Corporation and the host of the Infosmack podcast. Mr. Knieriemen took exception to statements about storage single sourcing I made to the press related to a recent document about storage choices for virtual server environments, including the following:
RT @markwojtasiak: Forrester says "single source when possible" <-what an incredibly dumb thing for @reichmanIT to say
He followed up his comments with the following clarifications:
@reichmanIT In context to managing data center costs, single sourcing is probably the worst suggestion I've heard from an analyst
@reichmanIT In a virtual server environment, storage is a commodity and should be purchased as such to control costs
I don’t know Mr. Knieriemen personally, and I must admit that I was taken aback by his blunt approach, but I’m from New York and have a thick skin, so I can look past that. The reason I bring this up on Forrester’s blog is solely to take a look at the actual points of view brought out in the exchange. The crux of the argument is whether or not storage is commoditized, and in my opinion, it’s not (yet anyway). There are three reasons why:
Yesterday, NetApp announced the acquisition of the Engenio storage division of LSI for $480 million. Engenio is mainly known as an OEM provider of storage systems with a broad partner list that includes IBM, Sun, Blue-Arc, Teradata, Panasas, RAID Inc., SGI and Huawei.
This move follows a busy year of storage acquisitions with HP, EMC and Dell each spending more than a billion dollars on buys in the space. $480 million represents the biggest acquisition ever for NetApp, more than they spent this past year on Bycast and Akorri, or previously on Onaro, Topio, Decru or Spinnaker.
Most of the money in storage acquisitions has gone to software that can be mated to industry standard (or nearly standard) hardware, but this deal goes in a different direction with LSI being a vendor known for meat and potatoes Fibre Channel storage hardware without a lot of frills. NetApp is largely a software company that sells OEM hardware running their DataONTap operating system, chock full of some of the most features and protocol options in the industry. Focusing on differentiated software has allowed NetApp to enjoy high margins and a consistent, unified family of products from entry level to enterprise class. So why would NetApp want to jump into the fairly commoditized storage hardware business?
This week saw one of the last remaining independent storage vendors (Compellent) get swallowed by one of the IT infrastructure mega-vendors (Dell) in an ongoing drive for comprehensive solution sets. We’ve seen a great deal of industry consolidation in storage, as vendors want to offer broad solution sets to buyers that want a single throat to choke, more financial stability than little guys can offer, and better integration of solutions.
It’s no surprise that Dell is bidding on independent storage vendor Compellent as they are still licking their wounds from the loss of 3PAR in a head to head megavendor bidding war which ended with HP winning (?) at $2.4BB. Dell announced today that they have offered $27.50 per CML share, which equals around $876MM, and represents a discount of around 17% to the $33 CML was trading at when Dell first made the offer (increased significantly over the past couple of weeks by speculators anticipating such a deal). The discount is a surprise -- 3PAR and Isilon (bought by EMC for $2.2BB in mid November) had similar revenues, employee counts and customer counts, so I had thought that the $2BB mark was the going rate for established independent storage firms, but then, I’m not a financial analyst, so maybe I’m missing something there.
What I do know is that it is hard for Dell to say in September that they are willing to spend $2BB+ on a storage vendor, and then when they don’t win, to say that the EMC partnership and their EqualLogic products is all they need. EqualLogic has been a great product in the SMB end of the market and a big revenue generator, but hasn’t moved Dell into the realm of enterprise IT leadership that they so crave. The EMC deal has limited margin for Dell as a partner, and doesn’t establish them as a visionary provider with the chops to solve big enterprise problems. So, it’s no surprise that they felt they had to do something.
In a recent blog post I wrote as a summary of the goings-on from Oracle OpenWorld this year, I said, "Oracle needs to enable consolidation of multiple applications on Exadata instances to make it more useful." Since that post, Oracle has contacted me stating that they, in fact, do currently enable consolidation of multiple applications within a single Exadata instance. According to Oracle:
"Based on what I am hearing from production customers, I believe that most or more likely all OLTP customers are running more than one application on an Exadata Database Machine in order to get better system utilization. Even the smaller quarter rack is simply too powerful for a single application. XXXX, XXXX, XXXX XXXX, and XXXX are examples of customers having multiple applications on a single Exadata Database Machine. This InformationWeek article, "Oracle Exadata V2 Customer Tells All," discusses how BNP moved 35 applications to an Exadata half rack.
Note that Exadata is based on proven Oracle RAC technology, which has been used to consolidate multiple applications to a single RAC cluster since 2001.
We also have customers that are running DW and OLTP on the same Database Machine, including:
XXXX XXXX – runs DW and OLTP on the same Database Machine “Deployed Oracle Exadata Half Rack to improve data warehousing and online transaction processing, which is essential to factoring interest rates and processing loans.”
EMC today moved to fill the hole in their portfolio in scale out file storage (and limited success in NAS in general) with the acquisition of Isilon, for a cool $2.25 billion. Looks like the HP helped set the going price for the few remaining independent storage vendors on the scene when they spent around $2.5B on 3PAR earlier this year. Isilon, similar to 3PAR, has about 1,500 customers and around $150MM revenue, so this is an acquisition based on technology and future growth potential, more than market share.
File data is often estimated as the fastest-growing data type and is also unpredictable in its growth patterns. Isilon’s technology uses a scale-out architecture that fits the behavior pattern of big users of file data. You can start small, with a cluster of a few nodes, and then grow over time — and you aren’t locked in to a given level of performance — you can deploy nodes with different performance and density profiles. This flexibility, combined with extremely easy management and refreshes without complex data migration or downtime, makes for a powerful combination. EMC will take a big step up from their current offerings with the addition of the Isilon technology, probably the best option from an independent vendor today.
It’s been a few weeks since the extravaganza that was Oracle Open World 2010, and I wanted to post a few words about what I came away thinking. Aside from the genuine America's Cup guarded by large men with earphones and white gloves, BMX stunts on closed San Francisco streets, Oktoberfest and the Black Eyed Peas, Don Henley and The Steve Miller Band, I’d have to call the event, well, slightly understated. Admittedly, I don’t know much about the world of middleware, so the big ticket product launch of Exalogic was largely lost on me, although it may well be a huge deal to those in the know. There was drama with Mark Hurd’s appearance after his recent hiring; Larry’s never been one to avoid stepping on toes so he was willing to antagonize a huge partner that shared the stage on his keynote. This Oracle/HP friction is an unavoidable reality of the era of co-opetition as Oracle transitions from being a software company that partners with hardware companies to becoming a one stop shop and a competitor to those partners. In fact it’s pretty similar to the dissing of NetApp, EMC and HDS that Larry doled out last year when he announced Exadata, but HP looked better in that round as they were the hardware partner before the Sun acquisition. Did you think Larry was going to become a hardware vendor without burning (or at least heating up) some bridges with partners?
Look past the hoopla and what struck me was the laser focus that Oracle is showing as they transform themselves into a true IT mega-vendor with the world’s leading database at its core. Here are some of the most striking elements of this as I see it: