Setting back network-based computing one return at a time

There are many advantages to running applications on a network, like not having to worry about an operating system, eliminating your local host footprint, etc. But there's a big unspoken but widely understood issue: it's trust. You need to be able to trust that things will work just as well on the network as they would on your local server. You hand-over control and therefore accept some risk in exchange for the benefits. The overall function should net benefits to the customer.

Back in the early days of 'cloud computing' circa 1999 when we were running client-server apps like Exchange straight over public IP, customers were taking trade-offs in performance but conscientious ASPs more than delivered in terms of cost savings, availability and reliability. The last thing you wanted to do was make things on balance worse for the customer. And the ABSOLUTELY last thing you wanted to do was lose someones data even if it was 'just email'.

Well, fast foward a decade later. Here's a $3 billion so called leader in personal finance software, Intuit, with thier leading tax prep solution, Turbotax, trying to remain relevant within a growing market of online personal finance and related alternatives such as mint.com, wesabe.com , and others, many founded by Intuit alumni. These startups get that the key to the battle is in building trust. Well, let me tell you, if you are trying to build trust in an online business, you don't do it by losing someone's data.

I just got off a call with the Turbotax support line, because the help page directions I had been following to access a previous years return didn't seem to work. After spending 30 minutes waiting for an answer (this is actually the average hold time) and another 20+ on the phone with support, I'm told that the 2007 to 2008 conversion I had attempted, per the online help instructions, can only work once. If it didn't work the first time then your tax data has been deleted.

Never in my years of experience with various forms of online-based services would I ever have expected to encounter a system that would, by default, delete customer data. I wouldn't write an MRD for a program like that. I had never worked with an engineer that would code something like that. Nor do I recall a sys admin who would ever manage an environment where you couldn't recover data even if the program couldn't access it.

Not only is Intuit showing incompetance in application development, but by losing customer data they risk pulling down the entire industry. Whether you're doing online storage, CRM, ERP or security across the public cloud or within a private cloud, we absolutely must be stewards of the customer's data and ultimately thier trust. We strive to bring net benefits. But at the very least, even if you as a company are incapable of evolving yourself, even if you can't deliver customer value within an online model, then for the sake of those companies around you who will make it, please at least 'do no harm' and don't drag down the industry around you.

Technology Value vs. Doom and Gloom

Granted times are tough. But it’s gotten to the point where economic fear-mongering is beyond a popular political tool and has become a competitive weapon. It would be laughable when math-challenged vendors spread rumors of layoffs that exceed the competitor’s headcount. But layoffs, furloughs and slowdowns are not a laughing matter.

Luckily some companies remain focused on delivering valuable technology, esp. where it’s solving important problems, like how the new KMIP encryption standard will make organizational data more secure. Or they're innovating to reduce data center complexity, lower costs or save the environment. Some of us will remain focused on delivering new benefits and savings through technology -- hopefully customers are finding this helpful and maybe even hopeful.

Fabric Applications = Private Cloud computing

I've never been one to add to what is already some serious vendor over-hype, e.g. 'Cloud Computing', but then again I'm not one to rant against something either (regardless of how funny that can be). I'm hearing that some customers find the concept helpful. Cool. I know I find it helpful to review why these things have value to customers.

In my day job, we've been delivering what we call Fabric Applications for many years. In concept Fabric Applications are similar to what is being called a Private Cloud, especially as it relates to fabric apps such as Storage Virtualization. Our company's focus up to this point has been specific to Storage networks (vs. LANs), though this will change with FCoE and CEE.

And I had been wondering if given the current economy, or trends in vendor consolidation, or a new administration, or Jessica Simpon's weight challenges, etc... if customers' needs were really changing with regard to Fabric Apps and this whole cloud computing thing. As you may recall, a key driver to network-resident computing is the value it brings to managing resources from many different vendors. As it turns out, this is still a growing need for companies.

Data recently released by a leading IT research firm shows that more that 80% of Enterprises are working with 2 or more storage vendors, and more than 25% of companies are now working with 6-10 different vendors - an increase over last year! In addition, enterprises that have products for Storage Virtualization are continuing to expand their use. The hype around ‘cloud computing’ and ‘private clouds’ is apparently there because there's value in it: companies want best-of-breed solutions along with management flexibility. And Fabric Apps or Private Clouds or Enterprise Clouds or insert-your-favorite-term-here, are the way to achieve it.

Bringing Clarity to Cloud Computing

There's a great deal of news out there on Cloud Computing. And a number of new innovators and recent deals show this to be an IT hot spot for '09. But as shown in earlier posts, the idea of running applications on a network fabric isn't new. Storage management apps have been running within the SAN fabric for sometime.

So, What fabric apps are on the market today?

If you're looking for a more efficient way to do Storage virtualization, Data replication or Data migration, to name a few, you can accomplish this Today with established, enterprise-class applications from brand name suppliers. Here's a focus on one:

EMC Invista -- this leading Storage Virtualization app runs on the SAN fabrics of hundreds of leading global companies, some with up to a Petabyte of data virtualized. Invista enables non-disruptive data movement, ILM, increases storage hardware utilization, supports server virtualization efforts, and does it all across a heterogeneous storage environment.

Invista has proven benefits: one company realized a 3:1 consolidation of storage equipment. Another saved 91% on storage provisioning time via pooling. And another reduced a 22-month migration plan to just 4 months. And being able to mix-and-match storage allows organizations to optimize on lower-cost hardware and reduce thier overall Cost/TB - a big deal in today's economy.

The hallmark of todays fabric apps is Enterprise-class functionality with very high-performance. If this is the space you're operating in, then you should definitely be looking into this quieter corner of cloud computing: SAN fabric applications.

Enjoy this very recent story on reducing storage cost through things like storage virtualization.

An old idea to get your brain around

I was swapping email with a pretty bright colleague of mine the other day, telling him about Fabric Applications, and he responded that he 'couldn't get his brain around it'. I assumed it was understandable given that Fabric apps are a fairly new idea to many. But just how new are Fabric apps? Not very - especially in IT-years.

With a little googling you can turn up a number of stories about the Brocade Fabric Application Platform, as well as Gartner's write-up announcing the company's entry into the market via the Rhapsody acqusition.

This was on November 8, 2002.

So turns out, Fabric Apps go back further than Camera phones and even VoIP.

The primary appeal of Fabric apps in 2002 was the same as it is today, to quote from a story at the time, "...to run [applications] in the SAN without the need for a seperate server."

For more info on this time-tested approach to enterprise-class data management that lowers administrative cost of ownership, visit the Brocade Fabric applications page.

Why Fabric Apps?

You may wonder, "Why Fabric Applications?" We already have servers, appliances and even storage arrays that can be used to run applications -- why another platform?

As it turns out, there's a huge demand for better, more intelligent management of organizations' data: whether to more optimally store it, more dynamically redeploy it, do more with it, better secure it, lower the cost per gigabyte or the required energy to support it. Running these data management applications 'on the fabric' turns out to be more effective, efficient and cheaper. Specifically:

• Broad OS Support -- Storage fabrics aren't tied to server operating systems, so they can easily work across heterogeneous server environments.
• Heterogeneous Storage Support -- Similarly, fabric apps work across vendors' arrays. This comes in very handy for lease expirations or any time you need to move between different vendors' storage equipment.
• Eliminating Host Footprint -- Fabric apps don't require server agents - this is a huge area of savings for large enterprises who would otherwise have to touch hundreds or thousands of servers, not just for the deployment, but every patch, upgrade, etc.
• Leverage for your SAN Infrastructure -- most companies have a SAN, and this is a way to leverage your current investment rather than deploy additional servers or appliances to support a new application. You also leverage your existing staff's expertise in SAN administration.
• Enterprise Scalability -- Fabric apps are designed to scale with the SAN - to thousands of hosts and Storage Arrays.
• Highest Performance -- Fabric apps run on purpose-built hardware rather than generic x86 boxes. This means that applications execute at wire-speed, on devices capable of millions of IOPS.
• Enterprise Reliability -- Fabric apps are using the same trusted data backbone that deliver 99.999% reliability in the data-path day in and day out.
  

Today there are a growing number of applications finding their way onto storage fabrics. An upcoming topic will be about some of these types of Fabric apps.

Fabric Applications - Welcome to the Fabric

Welcome to the inagural post of the Fabric Applications blog. What are Fabric Applications? The term encompasses data management programs that run within a SAN. It's derived from the Fabric Application Interface Standard or FAIS, which is a common API framework for implementing storage applications in a storage networking environment. More info can be found on the FAIS project site.

This blog is intended to become a mecca for all things Fabric Apps -- I hope you'll come here to read, contribute and help grow the shared knowledge of this new and valuable technology.