Data Center Strategies

NetApp does not buy Data Domain, EMC does

The last time I blogged on the Data Domain play, the title was:

“NetApp buys Data Domain” (5/20/2009)

So much for that.  Here’s the 7/09/2009 update:

Assuming no other suitor shows up at the altar, it looks like EMC has bought the Data Domain bride.

I have to wonder though – what does EMC get? One should be careful what they wish for. As some have said privately or publicly, is Data Domain really worth $33.50 a share?

Let’s step back for a moment.

Data Domain’s beauty comes from deduplication technology and of course, its broad customer set. The hardware that the dedup algorithms run on is not particularly unique, but does offer a means to generate profit margins. And, no doubt the service contracts are lucrative.

A few years ago, the ability to perform dedup duties was important as the technology emerged from obscurity. But now – my gosh – dedup is near commodity status. In fact, vendors are  starting to consider pricing dedup as a commodity; look at what Acronis is doing with their dedup pricing. Yes, they are small, but their dedup per unit pricing is a harbinger of the future.  

The dedup game has moved past who has the best dedup ratio to who can scale and best fit into large and complex topologies. Topologies encountered in datacenters and across remote offices. My bet is that dedup becomes a feature, expected as part of any storage infrastructure. Look at the message that Symantec is espousing – dedup everywhere. Everywhere means commodity. Dedup becomes a feature not a product. In fact, look around the backup and archive product offerings in the enterprise/SMB market. It’s hard to find a product that does not include dedup.

So I wonder what EMC sees in Data Domain?

No doubt it’s the customer base. That’s no small thing. Getting the proverbial EMC (and their partner DELL) foot in the door for every account is a big positive. For where there is a need for dedup, there is also a bucket full of primary data needing its own equipment with a dependency on dedup for backup space efficiency. In a conspiratorial way, EMC also keeps Data Domain out of NetApp’s hands.

But EMC already has dedup’ing equipment. With the Data Domain acquisition, confusion will reign across the rest of the EMC dedup’ing products. What happens to the EMC/Quantum based dedup products? I guess they get put on an end-of-life cycle.  Take note: dedup algorithms are not interoperable. Now EMC has a cacophony of dedup engines (Quantum’s, Avamar’s, and Data Domain) to mash together. Good luck with that.

Challenging hardware implementations, is a growing trend to move dedup capability into the software and thus divorcing it from the hardware. Just listen to Symantec and Commvault. They will be happy to fill you in. To meet this threat, Data Domain must separate the dedup engine from dependence on the hardware, but that kills the hardware revenue stream.

So in the end I’m confused by EMC’s move.

Just what is this new relationship about? Why is it worth so much? Data Domain employees and stock owners get a nice dowry but does EMC gets its money’s worth long term? The long-term prospects for profits from Data Domain dedup technology are, to be polite, murky. Short term, EMC can capture some net market share but marriages should be about the long haul.

Maybe NetApp is the lucky one.

They could conceivable pickup Quantum and its dedup technology for pocket change compared to the Data Domain deal. They won’t get Data Domain’s global customer base but they also don’t need to dilute their company’s value either. In any case, NetApp is certainly free to find a new suitor or suitee; but that’s a different story….

Posted by Gene Ruth

Citrix Synergy Part 1 - Keynotes

For those of you who didn’t attend Citrix Synergy, I’ve summarized the keynotes each day.  I’ve not put them in chronological order to include the exciting stuff up front.

Citrix announced a number of new products soon to be made available across all their solution lines.

XenApp:  CEO Mark Templeton introduced a new self-service application deployment system called “Dazzel.”  IT administrators publish available applications to a web-based interface so that end users can simply click on the application icon they would like streamed to their virtual environment.  It includes a “star field” view akin to George Lucus’ Starwars episode introduction screens – application icons scrolling into outerspace.  The concept isn’t new (Symantec Altiris, Novell ZenWorks, LanDesk, etc. have a similar concept), but the solution is integrated with the Citrix XenApp delivery solutions.   Additionally, Reciever was introduced – a light weight client for accessing XenApp services.  This includes yet another iPhone App that is now available on the Apple iPhone App store.

XenServer:  On Day 2, Wes Wassen (Citrix chief marketing officer)  introduced XenServer 5.5.  Re-iterating the XenServer 5.0 stratification change to offer a Xenserver 5 with live migration for free.  XenServer 5.5 will add a consolidated backup infrastructure, directory services integration, dynamic load balancing, and improved reporting – all of which are required to meet Burton Group’s server virtualization hypervisor requirements.  Citrix also introduced a new certification program,  Open Storage Program – to certify third-party storage arrays with both Hyper-v and XenServer through Essentials.  Stage management (not shown) will also be included in 5.5 for either Hyper-v or XenServer.  Labmanager, available in Essentials for XenServer 5.0, can select resources for lab testing from Dazzle. 

XenClient:  Ian Pratt showed off Citrix client vision with a demo of the new dynamic client with Intel.  Both companies worked together to move Xen to client environment.  Work started in 2007, issues with power mgnt and all the client devices (3D graphics) have been overcome with the Xen type I hypervisor below 10MB.  The demo showed two VMs on the laptop, one is personal (Vista), and one is business (XP).  They showed full Aero experience of Vista.  Demoed support of hardware policy (only certain USB sticks, or lock down completely).  Integration with Dazzle to get the apps the users want.  Security feature demoed with a keylogger running, an app deployed through Dazzle is marked as secure, when launched it comes up within its own special windows (exposed from a private VM) inside of the insecure desktop environment (the window is in a thin a green box).  Keyloggers can't get to this secured app even though it appears to be running inside the insecure personal VM.  Mac users running Parallels will recognize this as similar to Parallels Coherence – a window from a VM running inside the active desktop, but in this case, the active desktop is a different VM on the machine.  The XenClient is planned to ship this by year's end.  Ian demoed streaming windows apps to a MacBook Pro OSx machine that was running XenClient.  No timeframe for this product was given.

NetScaler:  Wes Wasson announced NetScaler VPX product  - a virtualized appliance for XenServer for any hardware.  Customers can download XenServer for free and then add the preconfigured VPX appliance and associate it with individual virtualized applications.  This brings multi-tenancy to a new level.  Citrix threw out a challenge:  $10,000 award for the person/company to show the best innovative idea with VPX.    The Tech Preview of VPX will be available May 18th with general availability in Q3.

Cloud:  Wes and Mark both discussed cloud directions by adding to Citrix C3 Cloud.  Multi-tenency with VPX, virtual switching with XenServer, enable apps and desktops as a service, and adding pay as you go.  Citrix is partnering with Amazon webservices to allow C3 to interact with EC2.  Citrix announced a C3 lab integrated with EC2 compute and storage to include services of Citrix C3 blueprints  - best practices on how to do a public/private cloud environment with Citrix C3 being the private side integrating with Amazon EC2 as the public side.  Both Amazon and Citrix sharing the guidelines on how to do this as part of the partnership. Mark showed a list of the top SaaS vendors:  Citrix has achieved #4 with gotomeeting.  Over 1 billion connection minutes in gotomeeting in 2008.  Larger than Microsoft Livemeeting.  Salesforce is #1 still.  Mark discussed Citrix view of cloud economics:  Economics then flexibility, automation, and finally pay-go.  automation is self service, pay-go is external cloud connected for burst over, DR, and lab/test.

Partners:  Patrick Gelsinger of Intel introduced the Xeon 5500 (Nehalem)  He showed doubling of performance over the previous generation.  30 performance records broken.  168% virtualization performance improvement.  They delivered it with 50% less power consumption.  Intel calls the 5500: "Technology foundation for the dynamic data center"  Adding to the 5500 are SSDs, and 10Gb Ethernet - calling this combination the cloud technology of tomorrow.  Demoed 16 four-year-old XenApp servers consolidated on one XenServer virtualized 5500.

The conference was kicked off with a guest speaker on the Xen of Presentation: Garr Reynolds (associate professor of management at Kansai Gaidai University, Japan).  Garr talked about how to simplify presentations and user interfaces.

The keynotes concluded with a CIO panel lead by John Gallant of Network World.  It included CIOs from:  Arizona State University, Deloitte, and Credit Suisse.  Mostly the panelists extolled the virtues of app virtualization and re-iterated the importance of keeping costs down.  Practical advice was to set expectations of what you will get and what the business needs - keep from increasing costs.  Credit Suisse had increased productivity by allowing remote access for users after hours but still allowing some level of user customization.

In part two, I’ll talk about breakout sessions.

[Posted by Richard Jones]

Cloud Storage: What we need now.

In my last blog, I talked about what cloud storage is, isn’t, and should be.

This blog is about what actions the industry should take. So here goes:

It’s time for the cloud storage industry to organize in a meaningful way.

Here’s what we need: A well subscribed cloud storage industry association. I’m not just talking about the many wannabe’s trying to grab a foothold. Those folks are needed of course, but big players such as Amazon, Google, Microsoft, and IBM need to lead. Some vendors will be service providers; others will supply the infrastructure necessary for building cloud storage. We need them all. Can SNIA do this? They recently announced a technical working group to focus on cloud storage. They quote my position. I hope it works out.

Cloud computing folks are trying to organize, but so far with limited success. This comes from a blog about the “Cloud Computing Interoperability Forum”:

“There is not and has never been an agreed upon definition of the CCIF. As organizers we have “announced,” at various times, conflicting statements on how “our members” should view this Forum. These definitions range from “cloud advocacy group,” which implies membership and organized offline activity, to the much narrower “email discussion group.” Due to our failure to better define our project each community member has been left to his or her own devices, latching onto any number of definitions.”

Hmm, good luck with that.

Any cloud storage efforts need to be more than defining API’s. They need to address bigger industry issues. Here’s what I suggest should be tackled:

First and for most, the industry needs a definition of what cloud storage is. How do you know it when you see it? There’s confusion on this point. The industry needs common terminology and a list of characteristics for cloud storage. A reference architecture would go a long way to defining what’s what. That should spur some heady discussions and a brawl or two. Nevertheless, we need it.

To help out, the Burton group will tread into the mist and release cloud research for our clients in the coming weeks. We address both cloud compute and storage, layout definitions, architectures and what to do next. Should be fun.

Second, divide the room into those providing cloud storage services and those providing cloud storage infrastructure. Do a Google (MSN, yahoo) search for “cloud storage”. Lots of hits. Like “Green” storage, storage vendors love to lace “cloud” into their marketing. It takes time just to figure out what a vendor means by “cloud” and if they are a service provider or what. This confusion is not necessary. The service and infrastructure groups will naturally have different agendas – let’s not mix them up and confuse things more. We need a taxonomy.

Third, start working on a management model. How in the world are users supposed to verify an SLA is being met? They can’t, at least not without writing their own monitoring application. Who wants to do that? The service providers need an equalized way to meter their services, monitor SLA compliance, and assign metadata. Some folks at the Open Grid Forum (OGF) have toyed with a machine-readable SLA proposal. Maybe apply that to storage.

Just those three things would be a good start, but it can’t take too long. Without common ground, the industry will die from individualism. Freedom to innovate is key, but too much anarchy will be counterproductive.

Oh, and yeah, let’s get an equalized API or two. Can’t hurt to have some common interfaces. While this may be the least of the industry problems and we need to be careful to not squelch innovation, at least this gives a chance for third parties to build services around cloud storage.

One final note: There is room for a middle tier of service brokers capable of connecting a user to a backend storage supplier, based on utility pricing and standardized interfaces. Enron gave that a shot – yes, Enron - obviously, they were ahead of their time. But, they may have been on the right track. Delivering interoperable cloud storage with a dependable SLA and pricing model will make it a commodity to a broker. Then perhaps a small part of Enron can be reborn and forgiven.

Posted by Gene Ruth

Cloud storage: confusion rains

Cloud storage is fogged up, the mist obscures it, bringing a veil of uncertainty. I feel the tension of an oncoming lightning storm. Hail pounds my roof, smashing my windows!

I love it. This whole cloud thing lends itself to word play.

Gets a little tiresome though.

I’ve been writing a research report on cloud storage for Burton Group clients and vacillate from being excited to discouraged about cloud storage. Many vendors offer it. Just ask. Do a Google search for “cloud storage”.

Lots of hits. Too many hits.

Sometimes I wonder if there is a contest to see how many times the word “cloud” can be used in marketing material. It’s similar to what happened when “green” storage came into vogue. Nearly every storage vendor has jumped on that bandwagon. Search “green storage”. Points will be granted for enterprise storage vendors who don’t mention green in their literature.

Not helpful.

Not helpful are all the vendors who claim some variant on “cloud storage”. Are they an infrastructure provider, a service provider, or a broker? Hard to tell without sifting through the marketing clutter. And where’s the business case? For large Enterprises or the SMB space, any contemplation of cloud storage should start with a ROI/TCO calculation or at least an attempt. Moving to cloud storage, whether internal or external is largely a financial decision. Can it save money and can an organization live with the tradeoffs? Vendors: we need some help.

So how do you know cloud storage when you see it?

Some folks liken cloud storage to a utility – like a utility, the data flows to a customer, metered in some way. A useful analogy, but not quite right. After all, taking the analogy further, it’s the lake above the dam that stores the data, eh, power, that gets delivered to one’s house.

A closer analogy is that of a physical storage unit – the kind that stores stuff that no longer fits in the garage. You rent a physical storage unit by size, it’s expected to be secure from theft, protected from destruction and the stuff in it available when needed. And when you no longer like the landlord or the rent gets too high, you can move your stuff to another physical storage vendor. A sore back may result but once moved you’ll get the same service. So, too, should be cloud storage.

OK, enough fun with analogies, but you get the idea.

Isn’t part of the value proposition for moving data to a cloud the blissfully ignorance of the machinations to store it? Now, I wouldn’t advocate using cloud storage if you are clueless of what the backend infrastructure is - but that's the concept.

Cloud storage will not succeed as long as vendors have their own spin on it. How could it? The idea is for users to be able to leverage cheap and reliable storage. Storage managed by somebody else with some reasonable service expectations. Sure service providers can offer premium and unique services but there must be a lowest common denominator across the industry. As a user I must be able to move my stuff – my data – between service providers at my whim. Maybe I want a better price, or the service provider is hard to deal with, whatever. As a user, particularly a large enterprise user, locking into a service provider gives away all my leverage.

It’s nothing more than signing a long term outsourcing contract.

That is not cloud storage.

I could grumble some more, but clearly if one looks into cloud storage at all they’ll see that the picture is muddled, foggy, misty, gray, ok I’ll stop.

Cloud storage needs some definition.

I’ll get into that in my next blog.

posted by Gene Ruth

 

Shhhh! Burton Group Website Easter Egg

Just in time for Easter, Burton Group has put an Easter Egg on the Catalyst Doorway page, offering a big discount for our Catalyst Conference coming this July.

If you've never been to Catalyst, then I highly recommend you go. As always, Catalyst is an intense, 3-day event that covers some of the most important topics in the IT industry. If you've been to Catalyst, then you know what I'm talking about....it's like drinking from an information fire hose. This year, I'm especially excited about our cloud and virtualization coverage. For cloud, we are launching our coverage in earnest (actually the launch is April 20th, another secret) by defining the cloud, discussing usage models, how to build an internal cloud, security, cloud economics, and pitfalls. For virtualization -- client, server and storage virtualization --  we'll delve into top-of-mind concerns such as storage connectivity, virtualization security, network connectivity and virtual switches, application development, new HW technologies, management....basically the whole enchilada.

As always, the DCS team (Richard Jones, Chris Wolf, Nik Simpson, Gene Ruth, and myself) will there to give our perspective and analysis on these topics, but we also have some killer speakers in Paul Maritz (CEO VMware), Mark Templeton (CEO Citrix), Bret Hartman (CTO, RSA), Simon Crosby (CTO, Citrix) there to discuss some of these important issues. 

OH -- and it's in San Diego again this year (my favorite Catalyst venue).

OK, to crack open the Easter Egg (and get the discount), here's what you do:

1. Go to the Catalyst home page (http://catalyst.burtongroup.com/).

Either: click and then drag your mouse off the logo and release the button. OR: roll over the San Diego button but do not click, wait about 20 sec.

2. A message will pop-up stating

"Congratulations! You’ve found an exclusive discount code for Catalyst 2009. Use promo code: Easter Egg and get General Sessions for only $999! Register today – this discount is limited to 50 users and could disappear at any time!"

3. Register.

That's it. Good luck and hope to see you at Catalyst.

[posted by: Drue Reeves]

SSDs: the techy world loves them

Take a look at the video on this link from Gizmodo:

http://i.gizmodo.com/5166798/24-solid-state-drives-open-all-of-microsoft-office-in-5-seconds

The folks in this video put together twenty-four flash SSDs to create a monster machine – pretty cool and amazingly fast. Faster than has ever been achievable on dorm room budgets. Well, a pretty nice dorm room considering the cost of twenty-four SSDs… maybe these guys drive BMWs….

Watching the video got me thinking – all this is exciting, but for the more reserved crowds working in data centers, it doesn’t do them much good. Except for the most extreme data center, maybe at a university, data centers will not be building their own SSD storage systems.

I’m hearing lots of announcements regarding SSDs but few system vendors are actually delivering anything.

Pillar Data recently announced the addition of SSDs to their systems. Pillar offers “application” aware configuration – a neat idea that uses quality of service to create storage elements. A nice fit for SSDs since, contrary to some thinking, SSDs are not good at everything. Great, but it looks like we’ll need to wait a few months to buy them. Same goes for Compellent – they’ve made some announcements and certainly have an advantage with the way they auto tier data blocks, but there is no wide scale release of products yet. So, unfortunately, the choice of enterprise class storage systems with SSDs is pretty slim.

But I want them now. The clients I talk to every day want them now.

Yes, one can buy enterprise storage systems with SSDs. For the fringe element, both TMS and Violin Memory offer turbo’d flash based transaction boxes. These are fast products to be sure but not for generic applications. For the generically inclined, one can pick up an EMC Clariion system with SSDs in place of HDDs. All well and good, but that’s not the game.

I’ve harped on this before.

We need systems fully integrating SSD capabilities – not just replacing HDDs one for one. Replacing HDDs with SSDs is storage design 101. Easy to do but very suboptimal. Blending an SSD into a system is the real challenge. For general purpose use, there is Sun’s Amber Road line (I like that name). A product that blends SSDs into a HDD storage system. Good for Sun but more is needed.

I’m hoping and guessing that the really interesting designs are being kept deep underground. Makes sense in order to protect secrets and patents. But who will surface first? FusionIO, with their PCIe based SSD is making lots of noise, TMS recently announced a similar product and is teaming with NetApp. Will IBM finally ship their much touted Quicksilver transaction machine? They’ve shown this system, built around their “SAN Volume Controller” at the Fall Storage Networking World. So where is it?

Sun keeps chugging along making announcements and delivering SSD product – see the recent announcement. Besides designing SSDs into their storage products, SSD can now be had in the server line. Excellent. And what’s up with the “Open Flash Module” they’ve mentioned? Hmm, let’s speculate….first, its a duh to design it into a server motherboard – goodbye boot disk – hello flash SSD boot. That should bump reliability and nick away some power. I expect to see this module show up in “mechanically optimized” storage equipment. Imagine the cost savings possible by removing all the cables, power supplies and enclosures needed for conventional HDDs. Making the flash module design “open” is interesting but what will it mean without taking it to a standard’s body? Probably not much unless Sun is planning to become a SSD module supplier. Possible but not likely.

And one last rant– we need more tier one disk vendors to enter the game to make the industry go. Not unlike the number of vendors we see delivering performance HDDs. More SSD vendors equals more competition, more innovation. With competition, SSD performance will be driven beyond what we are seeing now. While the performance of enterprise class SSDs such as Intel’s are impressive, I suspect that that this is just the start.  Keep an eye out for better write performance on par with read and less sensitivity to wear.

Expect more experiments like the one at the top of this blog. The computer community hungers for SSDs and is getting impatient.

Posted by Gene Ruth

Create your own stimulus package.

The constant bombardment of gloom and doom about the economy can wear you down. With such negative vibes coming our way, it would be easy to put those IT storage projects on hold, hunker down and beautify the cabling on equipment racks. I have no doubt that many IT groups are meeting to cut spending and minimize expenses that are not critically needed. Those in the organization who are good at justifying what they have and need will fare well. The less forceful or articulate, no matter the intrinsic value of their cause will suffer. And so it will go for awhile.

But this downturn will not last.

The economy will turn around.

We will not end up living in our cars. To think otherwise is foolish.

There is talk in the stock market that for those with iron fortitude, some would say brass you know what's, now is an historic time to buy stocks. Warren Buffet thinks so. I'm not sure - don't take investment advice from me, I still work for a living - but it makes you think: is now a good time to be building out new storage infrastructure? Will pricing on storage infrastructure be lower than it is right now?

Being an optimist, my sense is that storage vendors are desperate to meet sales goals - even reduced sales goals. I hate to be a shark and kick good vendors when they are down but now may be a great time to buy new storage gear. Resetting those pricey service contracts may also be interesting. After all, unless you've completely lost hope, it can only be up from here, right?

Sooner or later the economy will start to grow again. That means storage vendors will be taking orders. Vendors who no doubt downsized, will be constrained by a lack of production capability. Once things turn around, prices will rise as demand rises. If not from a constrained production chain, then from the zillions of dollars that are being pumped into the financial system.

The value of a dollar for computer storage may be peaking now.

If you have any faith in your business and the continued need for storage in your data center, then the allocation of investment dollars to storage will be a good bet. Updating storage gear to leverage thin provisioning, improve power efficiencies or reduce data footprints with deduplication can justify an investment. For those driving costs down with server virtualization, updating the storage to go with it is inevitable.

It's hard not to imagine a business that doesn't think it will come out the other side of this economic crisis alive. And if the business is alive on the other side of this gulf, investing earlier in storage gear for a data center will have been a good bet for positioning the business for growth.

So, if you've got any cash left, or even if you don't - vendors may be willing to finance you - consider updating storage infrastructure now. Create a custom built stimulus package. Being ready to handle an improved business climate will pay off in the long run.

posted by Gene Ruth

Linux, Where's Your VSS?

Its been nearly six years since Microsoft released Windows Server 2003 and its accompanying Volume Shadow Copy Service (VSS).  Most people think of VSS as Microsoft’s storage volume snapshot.  Yes it includes snapshot, but it’s much more than just snapshot, it is a snapshot framework that includes a rich set of APIs to allow third-party software integration.  This enables storage array vendors such as EMC or Network Appliance the ability to integrate their array based snapshot solutions with VSS.  It also enables application vendors the ability to construct their applications to participate in the snapshot process so that the volume snapshot of that application’s data is a consistent image.   Here’s an architectural diagram of VSS (courtesy Microsoft):

 

The framework of VSS is the most important aspect of this feature.  It forms the basis for enabling data protection using server-less backup.  In the architectural diagram, applications are the “Writers” and server-less backup software or snapshot management consoles are the “Requestor.”  Microsoft includes providers for its storage subsystem and storage vendors can plug-in their own providers into the framework.

The nice thing about VSS is that applications, especially databases, are given a signal to make their data in storage consistent.  They can flush buffers and transactions as if they were cleanly shutting down and then signal back that they are ready.  The snapshot then gets a good looking image of the application’s data, one that the application knows it can easily access without requiring lengthy repairs or playing re-do logs.
So now let’s go back to the beginning of my blog.  It’s been six years since Microsoft released this feature.  Storage demands have continued growing and don’t look to stop anytime soon.  The number of server-less backup products available for Windows Server continues to grow and enterprises have been deploying these products over the past couple of years to shorten the dreaded “backup window.” 

But where’s Linux?
Linux distributions don’t have any equivalent feature.  Yes, modern Linux distributions include the dm-snapshot module, but that just does the snapshot, there is no framework for application developers, storage vendors, or backup vendors to integrate with.  The result has been a few one-off solutions focused on a particular application stack, but for the most part applications just suffer from inconsistent snapshots on Linux.  These are what we call crash consistent, meaning that the application is responsible for figuring out the state of its data within the snapshot and getting it to a position it can begin running.  Most databases can take upwards of twenty minutes to mount a crash consistent snapshot.

So this is a call to the Linux community:  Please band together and architect a snapshot framework for all Linux application developers, storage developers, and users to benefit from.

[Posted by: Richard Jones]

May You Live in Interesting Times!

I’ll admit it - I’m extremely envious of our storage analyst - Gene Ruth. I’ve heard our analysts call storage "borage" because it just didn’t change that much in the past decades.  Yet SSDs have emerged as a (the?) disruptor that the enterprise storage market hasn’t seen in years, making us all envious of Gene’s job. 

Outside of work I’m a hardware enthusiast. I buy and build systems not because they’re needed, but because I relish the whiff of solvent that comes with the opening of an anti-static bag. Newegg (www.newegg.com) has been selling consumer-level SSD devices from OCZ, Patriot, Kingston, and others for several years, long before we saw serious enterprise offerings from Intel, Hitachi, and Samsung. It’s amazing to see consumer-level vendors announce an SSD, sell it for ~4-6 months, and then announce a replacement model with increased performance at reduced cost. I’ve been lurking, waiting for that right moment when price drops and performance increases. Along the way I’ve learned quite a bit about SSD OS tuning and benchmark limitations. Simply stated, there are alpha geeks among us who spend more time tweaking their SSD than actually storing files, and who are obsessed with finding the optimal block size or registry setting to extract the most from their investment. For example, when OCZ users were puzzled by strange benchmark performance results, they dug deep - just take a look at this post.

Today I came across this review of ACard’s ANS-9010 Serial ATA RAM disk. The ANS-9010 is a 5.25" device with 8 240-pin DDR2 slots, two SATA 3 connectors, a small battery, and a CF slot for backup. Stuff it full of your favorite RAM ($100 per pair of 4GB), and you have a SSD that is as fast or faster than the Intel X25-E Extreme SSD - in some cases by 2x, especially on large files or simultaneous operations - something that’s harder to do on a flash-based SSD like the X25. 

My first impression was that this was another Project Pluto - something that technically could be built, but probably shouldn’t as it’s inherently dangerous. Who would store their data on a volatile storage device?  Yes, it has a push-button backup/retrieve to the built-in CF slot, but the lithium battery doesn’t have much juice beyond the time it takes to backup the drive. And what are the chances that someone is standing by the device during a power failure, ready to push the backup button? Why would someone spend more money building a device that, overall, is about as fast as an Intel X25-E?

But I realized that I’m missing the point. We’re experiencing tremendous innovation and rapid evolution in SSD devices. It’s highly unlikely that any of our enterprise clients will pitch this device to their boss. But it’s interesting to see innovation around DDR desktop memory and dual SATA 3 connectors for built-in RAID0. ACard has managed, with a bit of duct tape and bailing wire, to outrun Intel’s best SSD on certain benchmarks. I know my fellow hardware enthusiasts will be tweaking this drive to squeeze out every last bit of performance. And who knows, maybe those lessons will appear in a future enterprise devices. Stay tuned.

Putting Your Data Center on a Diet

So Christmas is over and we are all looking somewhat ruefully at our waist lines and hoping to get back in shape before the summer rolls around. Data center operators have another problem, how to cut some of the fat out of their data centers and get their energy budget in shape for the new year. Just this week I was asked the question “How can I trim a million KW-hours energy consumption this year?” That sounds like a daunting, but worthy exercise, a million KW-hours sounds like a huge number, but at 8-cents a kilowatt-hour trimming 1 million kilowatt-hours would save $80,000 dollars in the first year (of course, price/KW-hour varies considerably by location, so your mileage may vary.)

The first point to note is that while the number seems huge, that simply reflects the size of the customer’s data center, the real question is “what changes will have the most impact on energy consumption?” So here’s some tips on how you can reduce energy consumption in the data center:

1. Reduce the amount of equipment in use: The quickest way to significant savings is to cut back the amount of equipment in use through technologies such as server virtualization. Let's assume that 1/3rd of the servers in a data center are suitable candidates for virtualization and that on average a 5:1 consolidation ratio is possible (and that’s a conservative estimate for the consolidation ratio.) A datacenter with a thousand servers would be able to virtualize 333 of them, replacing those 333 physical servers with just 67, a net reduction of  266 physical servers. Now lets assume that these original servers were 1U pizza-box servers each consuming 300 watts of power, and that the new servers are bigger (to accommodate the extra memory, processors etc) and consume 600 watts of power, the net power saving is almost 60 kilowatts. Such an approach would also reduce power consumption for cooling by a similar margin (assuming 1:1: ratio between power consumed by equipment and power consumed by cooling.) This step alone would save a million KW-hours, throw in some improved storage management practices and deduplication for data and the energy savings quickly pile up.
2. Get rid of old equipment: Some data centers continue to use equipment until it finally dies, that can mean servers that are 5 or even 10 years old (I've seen even worse cases). These servers are energy and space hogs and deliver very little work for the power consumed compared to a more modern system, especially if the more modern system is combined with server virtualization. 
3. Improve power distribution: This can take the form of a number of improvements and upgrades. For example old UPS equipment is substantially less efficient (i.e. wastes more energy) than modern UPS systems, replacing the old UPS could yield a 5-10% reduction in energy consumption. Also minimize the number of voltage conversions between the power as delivered by the utility and the power consumed by the equipment, i.e. bring high-voltage power as close to the racks as possible.  Data centers in the US still using 120V AC to power equipment in the racks, need to take a long hard look at switching to 240V AC, the equipment doesn’t care and it will save energy and simplify the wiring.
4. Cooling improvements: Air-side economizers can deliver big savings, especially in more temperate climates. Assume roughly half the energy consumed by the data center is used for cooling, and that the airside economizer can function for half the year, then potential savings in energy consumption are pretty obvious. Also improve hot-aisle/cold aisle separation or implement fully ducted hot-air removal from racks can increase cooling efficiency inside the data center.

The big lesson here is that the quickest payback for reducing energy consumption is to reduce the amount of equipment in the data center and opportunity presented by maturing server virtualization technology is too good to pass up! The downside is that to save money in the long term, you’ll have to spend some money in the short term.

Posted by: Nik Simpson