When you shop for a car or a house, there are basic requirements that you typically have in mind. If you have five kids, I would imagine you might consider a minivan or SUV, and certainly not a 2-door hatchback, or a one bedroom apartment. ..

A friend of mine asked me the other day if I had an Axiom at home for all my music and video content. I laughed because it’s not really a product designed for the home.  It’s too big, too powerful, and has way more capability than I need attached to my home 100 Mbps LAN.  It’s simply not a good fit – much like moving my 4 person, 2 weenerdog family into a 16 bedroom 20,000 square foot mansion; we’d be lost in the thing.  I think there are times my kids would like that – as we could look for them for hours and not find them. On the other hand, they could look for us for hours and not find us…..hmmm.

Anyway, it’s best to match your needs and requirements to the solution, right?  The smallest Axiom we sell would most probably weigh in at around 5TB (assuming I used SATA), so it sure it would be fun not to have to delete anything for a long, long time.  But who the heck wants to power up 13 spindles when two or three is enough?
There is another side to consider – growth.  Why buy a house that fits your current family if you plan to expand it?  Do you have to move, or can you add on to the house you own?

Pillar builds two products: the Axiom 300 that supports from 12 to 52 spindles and the Axiom 600 that supports 24 to 832 spindles. They overlap a bit, and span quite a range of capacities. The Axiom 600 can support from one to four Slammers (8 control units). We have customers that started with 26 spindles who now have over 500.  Customer growth is not unique to Pillar. Everyone has Customers who started at 20+ spindles and now have 500+, don’t they? Well yes… and no. Most probably not on one system, as most storage companies have platforms divided into far smaller granularities. Companies that start with 20 spindles and grow to 500+ either buy additional systems or upgrade between model numbers.  For me, I just find that level of granularity annoying. Let me illustrate my point with a table:

So if you look at these comparisons, Pillar covers the NAS, SAN, and iSCSI markets with two nearly identical platforms.  EMC has the CX4-family to cover SAN and can add a gateway (or just sell the NS4-family, which is also a gateway just bundled together with a CX4) to provide NAS functionality.  Upgrades between models, or adding the gateway to a CX4, is pretty damn expensive. And what’s with that big gap between the NS-120 and NS-240?? If you’re going to granularize your products why leave a gap?  Go all out!! NetApps also has major Software license issues when uplifting between each of their platforms.  They do offer SAN and NAS in one “head” but the issues that has caused are well known.  NetApps is a NAS company.  .  Then there’s Sun.  Their “SAN” systems (primarily LSI systems) don’t even have iSCSI.  Their NAS systems don’t have a FC SAN option.  Hardcore purpose-built storage (reminds me of the 90’s).  Frankly Sun has a bit of a problem: proliferation of models, lack of platform scalability and almost no Sun intellectual property.  They OEM the majority of the storage line from LSI and HDS.  Their own storage plays in the lower end.  It’s supposedly pretty good NAS stuff but is it “good storage” or “good” compared to Sun’s past crap (remember the T3?)?

I think of the Axiom as a modern day freight train – you attach the number of engines (Slammers) you need on the front of the train, and the amount of cars to hold the freight (Bricks) on to those engines to fit your weight (IOPS/bandwidth) and cargo (data capacity) requirements. You have building blocks to get the job done…and, you don’t have to change model numbers (platforms) and pay the massive fees associated every time you add an engine or a car to the train.

A few people have asked me what our corporate messaging really means. In fact, some of our competitors have even embraced our message as their own. I think it’s funny others are copying it – especially those that don’t even seem to know what the heck they’ve copied. At Pillar, we actually do care about what it really means, so I will explain it with specific examples.

The true value of Application Aware storage is the ability to set Application Profiles and configure hardware resources to improve utilization and performance. Profiles segregate LUNs or file systems within an application – each of which has its own Quality of Service (QoS).  The entire application performs at a higher level if this is done properly (which is what the profiles do automatically if you are working with a common application). The result is that the total application performance is greater than the performance individual data stores!  

For instance, in a Server Virtualization environment it is the segregation of the OS Volume, the application volume, and swap space.

In a Virtual Tape Library (VTL) environment, it is the VTL volume, the single instance volume, and the index volume.

If you follow EMC best practices, you put all your eggs in one basket by creating a virtual machine on each file system. But with Pillar Axiom storage system you segregate each virtual machine in to its components as I outlined above.

To illustrate, we configured an Axiom for three simple cases:

VM1: OS, application, and swap all on a single SATA LUN
VM2: OS, application, and swap all on a single FC LUN
VM3: OS and swap on separate SATA LUNs; application on FC LUN

We observe that VM3 performed much better than VM2 (as much as 2x) or VM1 (as much as 3x)! 

The beauty of it is, that you don’t even need FC disk drives. The above setting is just one profile – for one VM! If you have 2 VMs and the VMs need different QoS, you can do that too. 

Here is a profile of 2 VMs running similar applications in tandem and needing different Quality of Service (QoS) levels.

Now that is truly compelling.

The Axiom policy-based manager allows you to do this in about five clicks of a mouse. While it is possible to lay this out on a competitor’s system in a similar fashion, should you need to change it, all that work has to be done all over again….and you have to migrate the data!  Compare that to Axiom – where you just change the QoS, decide whether to allow data migration, and all the work is done for you. Hell, you can even change it automatically at the end of month crunch time for example and then change it back.

The above examples illustrate the power of a storage system invented more recently than OS/2 version 2 and Windows 3.1….(that would be circa 1992).

I have to say – except for the awful name, I am impressed with this thing!  Clearly a lot of slideware, but the intention is impressive. Before the Pillar lovers and EMC haters get all ticked off at me, let’s step through a few of their marketing talking points:

I suppose they saw a 25% drop in Symmetrix sales last quarter because their customers were waiting for this one, huh?

I am impressed that they’re copying (albeit only an intention so far) the Axiom and its key features.  It is kinda strange too, they haven’t replaced the Sym IV, but say that the new one does everything that the old one does…must not be a compatible upgrade…

Clearly EMC marketing wants to make this sound new and exciting.  But I have to question a few things. They call it “purpose built.” Purpose built? That’s so 1990’s!  The whole industry has moved away from that model. The Axiom was designed specifically NOT to be purpose-built. EMC has a history of this (the Centera and EDL come to mind). Even better, it’s purpose built for the virtual data center. Last time I checked, one of the biggest driving forces to implement virtualization is to decrease cost.  Buying a Sym certainly doesn’t jibe with that!

These are just a few of their “innovations” or marketing selling points. Frankly, there is no doubt that a) this announcement is way over-hyped, and b) it’s way overdue. As a good storage analyst told me “You have to give them credit though – their marketing team is unabashedly boisterous and aggressive.”

I am kinda at a loss for what else there is to say that hasn’t been said. While it may be revolutionary for EMC, it’s not for the industry.  I suppose us small guys can be happy that our ideas are eventually adopted by the big guys….but doesn’t that say something about being a “big guy”?

If you want innovation, ease of use, and the features a Pillar Axiom has now, look at the Pillar Axiom now, or EMC in 3-5 years.

Well, now just when the bloggers who have a day job thought there might be a lull in the action, Oracle pops up with its intention to purchase Sun. You have to write about this, right? Sure, everyone and their Grandma will….and Grandma probably has a blog these days, and you know the old bat is trying to scoop you.

Well here is what I have to say: Look out everyone else. This is a storm on the horizon that you don’t want to ignore.

I also have to say that I have no inside dope on this whatsoever, all my opinions are just that – opinions – and they are indeed mine.

When Larry started Oracle, he started with a paper that had been written by IBM on relational databases. Larry is the master of recognizing technology and turning it into a commercial success; Brains, technically savvy, and commercial know-how.

When I wrote the Blue Sun post, I was sad about the fact that IBM would most assuredly put the final stake through Sun’s heart. Sun has played a key part in ruining Sun up to now. Let’s be honest, it wasn’t Bill Gates. IBM cannot recognize key technologies anymore. They bought Storage XIV didn’t they? Besides, whenever IBM gets hold of a good technology it fumbles around with it until they are forced to sell it to the Chinese.

But Oracle on the other hand, whoa, these guys can turn a post-it note into a business! Oh wait, that was 3M. Anyway, Oracle will sort through the piles of great stuff (expletive deleted by me – aren’t you proud?) and turn it into gold. With Larry’s bold and aggressive personality and Safra’s huge operational focus, I think we might just have a winner here folks.

What will be interesting to see is how the coopetition unfolds…before this Oracle didn’t really compete with its partners, unless you count the database storage product that it did with HP. There are two ways to think about Oracle’s Exadata: 1) it’s too small to threaten any of its partners, or 2) it is too specific a solution to generally threaten anyone. Hmmm.

More importantly, I think this is good for Sun and their employees. Sun is a national treasure; a setting Sun is not good for anyone.

I was visiting a prospect the other day and she asked me what our “Sweet Spot” is. Well, I could give a marketing answer, or say “We sell into every industry, every Tier of storage so we don’t have one,” but that refuses to answer the question. So I had to think about it. You can answer that by industry, vertical, revenue, price range, all kinds of cuts at the question. All are valid and interesting to different people.

Here is how I answered the question: mixed tier, multi-tenant environments where consolidation and flexibility are of paramount importance.  As a CEO and technologist I’m tempted to keep throwing in more and more good stuff like power efficiency, multiprotocol, scalability, and all sorts of attributes that are part of our design, but she wanted to know where we have the largest, sweetest, most obvious payoff.

Why? Because we don’t do purpose-built storage. We don’t design stuff that must be “stove piped.” Sure we have customers using the product that way – hell maybe they replaced one company’s specific solution for “an application” with Pillar, and that’s all they wanted. We love it when they start throwing more stuff at the Axiom array – ‘cause it works. Quality of service doesn’t mean as much if you’re not balancing between multiple applications, application owners, and performance and capacity requirements.

We have lots of universities and museums as customers (lots for Pillar by proportion, nothing for EMC). These type of environments are full of multiple requirements, varying degrees of funding (and thus affordability of a given solution), and tough staffing constraints. Ease of use really pays off when a customer doesn’t have big bucks to spend on professional services to configure and manage storage. 

If you have seven different platforms and you want to keep them (maintain, license, and stay knowledgeable and proficient in each), Pillar doesn’t offer that. But if you want a platform that can be deployed against multiple requirements, a platform that is scalable and flexible, we have it.  You can always build a single purpose solution, say a digital security surveillance system, a VTL solution, or the likes on an Axiom and lots of our Customers do it. While I am confident that it is sweet for them, there is lots more in the Axiom for them to parlay into even bigger payoff than single focus applications.