LoudSpeakers

Sanjay Kumar, the former CEO of Computer Associates, was sentenced to 12 years in prison for his part in the accounting scandal that shaved billions of dollars from the company’s market cap. 

He plead guilty for his part in masterminding a 35–day accounting month that also allowed for software sales to occur during a given quarter, but allowed the customer to return products or not pay for the transaction.  In effect, the company’s performance was artificially inflated to achieve revenue targets.  Meanwhile, he personally profited from the company’s perceived strong performance. 

I wrote about WinFS late last year and provided some of my thoughts on the subject. In the days and weeks after writing the post, I noticed a massive number of referrers in my logs from microsoft.com directly to that post. For a period of weeks, it was my most viewed post — mostly from employees at Microsoft. Apparently, it touched a nerve and people must have received it in widely sent internal emails.

Microsoft announced recently that they’re essentially slowing the growth of the project by including it as a component of SQL Server. I have never fully liked the idea of an object-relational file system —- especially if it comes from Microsoft. That’s not because I enjoy Microsoft bashing, but I have real-world experience with storing massive numbers of files in Windows file systems.  NTFS just doesn’t cut it when you exceed 25 million files per partition. If a machine sets the dirty bit on that partition, a reboot can take days potentially to complete due to checkdsk. Adding layers to the file system to add nifty functionality is great in theory, but in practice it’s almost impossible to make it scale for enterprises when it’s based on a Windows file system.

WinFS sounded better for the consumer in some regards. They’ll never need the same sort of scalability that enterprise data centers demand, but the adoption of such functionality is still dependent on application developers seeing value in the offering.  Bundling WinFS with SQL Server is one assured way of keeping it out of the consumer space unless perhaps an application uses the developer edition of SQL Server.  But this is still a big barrier if Microsoft wanted to see adoption in the consumer space.

I spent the better part of the last decade dealing with issues surrounding adding metadata to NTFS using something called the extended attribute block. It’s sort of a free form area of the file system for low-level developers to store stuff. It was a bad idea to use the EA block to store meta information and I doubt that WinFS was much better.

EMC, my former employer, announced it was acquiring RSA for $2.1B. It’s their ballsiest (sp?) move to date. They’ve been talking about buying into the security wheel for sometime, but I thought it would continue around acqusitions like Enterprise DRM vendor Authentica.

I would suggest that acquisition of RSA is the most radical yet complimentary acquisition by EMC to date. Legato made perfect sense. VMWare was brilliant, but makes total sense. Documentum? 2nd most radical, but brilliantly executed.
After this acquisition, EMC is probably in top 5 largest software companies by revenue in the world. Not too bad for a company originally known for storage hardware. I’m curious as to who else was involved in the bidding war. Aparently, it was fierce.

But don’t forget about EMC as a services company. They have quietly been snapping up professional services firms in the recent months. This is a company determined to win the low-margin hardware game by combining them with high-margin software and services.

Update: This BusinessWeek article mentions that the other bidder was Symantec. This wouldn’t be the first time that Veritas/Symantec lost a key bidding war against EMC. See VMWare…

Amazon announced a storage offering for third-party developers to use. They will offer no client UI of their own, so it will essentially be “black box storage” for developers who would rather not invest in the expensive infrastructure involved in buying equipment and software.

Personally, I like it. They’ve reduced one of the larger barriers to entry for developers that need highly-available storage. Developers are no longer remanded to purchasing expensive storage systems while in startup mode without sacrificing the need for high availability.

There’s a ton of new services that can come out of it. It really does change the game considerably.

Removable media such as tape and CD media disintermediates the data from the application that created it and thus the ability to quickly identify physical losses such as the one that occured with Deloitte and Touche recently.

A Deloitte employee lost an unencrypted CD that contained personal information regarding McAfee employees and their company stock option plans. What’s worse is that it was left on an airplane — not destroyed in a server failure that would otherwise go unreported.

This is a huge problem for companies. Deloitte is just the most recent in the spat of firms that have suffered PR nightmares by losing removable media. Time Warner, Iron Mountain, UPS, and Citibank have suffered similar losses involving removable storage media and personal information.

Portability of media is indeed useful, but introduce the human element and big problems arise. Actually, it’s not just removable media that exhibits this problem — it’s backup in general.

Backup applications create “backup sets” that usually contain media of some type — usually tape media since it’s relatively cheap. More recently, companies are using disk based backup since the price/performance ratio is pretty good these days. But whether it’s actually removable media contained in the backup set or SAN attached disk makes very little difference after the backup process itself has occured.

Backup applications themselves usually know very little pertinent information about the data being managed. They only know about physical storage descriptions like a file system and possibly a database such as Oracle or SQL Server. Backup applications know nothing about the value of the data itself like the name, SSN, and other meta data that may relate to that data.

The backup set, once created and populated with data, no longer has any meaningful connection to the process-based application that created the data. It’s managed independently and thus creates a huge problem on many levels in all industries.

The examples of this are fluid and the problem is obvious, but the most appropriate solution is far from easy and likely impossible. Why is it impossible? That’s a subject for a couple of beers and a late night. Wait — I did that recently….

The root of this problem is 30 years old. The file system, in it’s modern form, is not capable of dealing with the requirements of today’s datacenter and the compliance challenges exerted on every industry and market. It has long been used as a cheap and loose integration point between process based applications used by business people and infrastructure apps in the datacenter. Ineffectual database architecture by the likes of Oracle and Microsoft just exacerbate the condition. (Pssst Microsoft: adding relational database features to the NTFS file system won’t solve this problem either unless developers take advantage of it.)

Separating data from the business logic in the process of management is a bad thing. The problem will only get worse.

I wrote recently about FreeNas, an open-source lightweight OS appliance based upon BSD and the m0n0wall project.

Before discovering FreeNas I came across another open-source project called OpenFiler. OpenFiler appears to be the special sauce behind the hardware developed by Xinit Systems. OpenFiler is based upon CentOS and has a much larger footprint than FreeNas though it is indeed aimed at being an appliance as well.

From the perspective of an enterprise deployment I like the concept of OpenFiler quite a bit. Choose your own hardware, possibly from Xinit and buy support from them as well. It’s not that different from the model of other enterprise open-source vendors.

My problem is that I can’t give you a simple review of OpenFiler because I couldn’t get very far with the install. As a matter of fact, it wouldn’t install at all on rather generic x86 hardware. It was an installer issue, but it prevented me from being able to get a glimpse of what the appliance offers.

For my purposes, paid support is out of the question and the particular bug that I encountered seemed to be acknowledged with no workaround or estimated fix.

Maybe you’ll have more luck than me. If you are able to get the latest OpenFiler stuff working, let me know.

Richard Morrell, the creator of the once uber-popular SmoothWall firewall distribution, is at work developing an open-source NAS solution called NetServa. He’s an interesting character for many reasons, but based upon his work with SmoothWall it’s obvious that NetServa is something to keep an eye out for.

I love turning old PCs into something useful. They’re perfect for appliances. They are slow as hell with Windows running on them, but they become supercharged machines when running stripped down versions of open source operating systems with specialized services.

I already run an open source firewall called m0n0wall. It’s one of the most stable pieces of equipment in my house. The uptime is only affected by a severe power outage. And it’s rock solid being that it runs BSD.

I have tried every other open source firewall out there, but m0n0wall is superior in functionality and overall operation. It even has an included VPN server so I can securely access all of my home files and network when I am on the road. It even supports wireless PCI cards so I can roll my own access point if I so choose.

A smart guy by the name of Olivier Cochard adapted the code from m0n0wall and created a NAS appliance product out of it called FreeNas. I just bought a 250 GB internal drive that was about $50 today for an old PC to run FreeNas. I’ll surely add more later as I test it out.

FreeNas is still taking shape so it’s not exactly easy for a novice user and it still lacks some important features that are found in commercial applications, but over time this will surely fall into place. But early iterations show that performance is quite respectable. I didn’t break out ethereal, but just eye-ballin’ it shows that it cranks along with massive file transfers.

So I can literally throw together a terabyte of NAS storage for about $200 since I already own the old PC and the software is open source.

I remember 10 years ago when a terabyte of storage was over a million bucks and now I can afford to put one in my closet as a blackbox network backup device.

The commoditization of hard disk drives has been overwhelming over the years. The result of this commoditization has also been a drag on the cottage industries that built products around the notion of expensive storage and how to deal with it better.

But this same drag has had an opposite effect on the consumer market. Go to Fry’s, Best Buy, or CompUSA and ask for a NAS appliance. This is something that you couldn’t really do 12-18 months ago. They might not have known quite what you were asking for then, but Netgear, Linksys, and Buffalo are just a few of the vendors manufacturing NAS appliances for the consumer space.

It makes total sense. Blackbox storage has benefits for the end user. A whole slew of communities has formed around the creation of replacement firmware for these devices as is also the case with video game consoles and wireless access points by the same vendors.

The only problem with blackbox devices for the majority of the consumer space is that they can only be mod’ed to a certain extent. I mean, that’s what blackbox devices or appliances are meant to do. I have a feeling users will want more than just storage for MP3s and such.

I have a feeling that these appliances will also drive the need for continuous data protection (CDP) software for the consumer space. CDP is essentially event drive snap shot software that lets you do a point in time recovery based on an event.

But the problem is that the NAS appliance would have to run some sort of CDP service and currently that is only limited to enterprise type of applications because of the costs involved.

Google is trying to get into the enterprise application market. They get it. Kinda. Well, not really. Indexing everything on the public internet is a great thing. Scaling it down for enterprises is just not the same thing. Enterprises have vastly different security practices that are employed when compared to the public internet. They also employ very different applications and have different objectives with their data. Over simplifying the scenario with something to the effect of “everyone needs search” and “unstructured data makes up for 80% of total data managed” is irresponsible and naive. In fact, implementing some combined technology from these companies could create a large liability for the customer because it may allow exposure of sensitive data once secured in a protected repository.

Google and StoredIQ are getting together. Makes total sense from a biz dev perspective. Looks good on quarterly MBOs. StoredIQ guy will be telling his grandchildren how he “did the Google deal.” I would hate the be the enterprise sales guy who has to go sell that when a technology guy like me is responsible for due dilligence. My next post in a week or two will be a list questions for a customer to ask of an ICM vendor when considering one of their solutions.

Great. Proves me point: Infrastructure vendors don’t get it.

Link: NetApp to OEM data classification software

What is Storage 2.0?

I promise that I’ll get off of the del.icio.us kick that I’ve been on lately. But the functionality employed by del.icio.us partially lends itself in analogical ways into what’s being called Storage 2.0. What is Storage 2.0? It’s the less geeky way to refer to Storage Existentialism. I also recently described it in my post regarding “The Star Alliance” of Storage.

With del.icio.us, users tag web pages and other objects for themselves and others. They’re creating user managed pointers to data that can be shared and reused.

The concept of user generated tagging already exists in the business world and it’s been around for a long time. Storage 2.0 is the new layer of data storage management that can leverage user and application tags. It refers to the way that business data will be tagged and managed by real-time value rather than by physical storage aspects. It’s not limited to unstructured content like Word, Excel, and PDF files. It spans all business assets such as radiology images in a hospital, video feeds in a Digital Asset Management application, and structured objects found in a relational database.

So the new layer that represents Storage 2.0 consists of reusable, programmatically generated tags that describes data for the purpose of more efficient storage management.

“How will WinFS work in the Storage 2.0 world, Raj?”

So you’ve been reading up on the next generation Windows relational object file system and you want to know how it will play out in the new layer of storage management. The real impact of WinFS will depend on adoption by application developers. Without being “aware” of WinFS, it will have no benefit to users other than the fact that Windows Explorer will have extensions that allow you to view a “document that represents other documents.”

But in the world of storage management for businesses, user workstation file systems have no implication on the data center. So how will WinFS work in the data center where Storage 2.0 really fits? To take advantage of this new functionality, server applications and processes will still need to be aware of the ability to create and read data in this new relational object file system. There are many server type applications that can benefit from such a file system, but developers have been overcoming this for years with other methods that are actually quite elegant and graceful.

When (and if) server application developers take advantage of the functionality in the new file system, it will essentially become a transparent part of the Storage 2.0 layer. This is assuming that enterprises choose to store data on a Windows managed file system. Enterprises aren’t ditching their high-powered, industrial strength Unix and Linux based storage servers just yet.

Microsoft will need to prove the scalability of the file system with some huge benchmarks before enterprises trust it for mission critical storage. Though the theoretical limit is very high, NTFS has been known to have real-world upper limits in terms of file count per partition due to performance issues.

I just read a brilliant post on Sandhill.com by Romesh Wadhwani (another one of my peeps) entitled “10 Opportunities For Software Vendors.” The point that really strikes a chord and relates to my apparent ongoing rant about monolithic software solutions is:

The Emergence of a “Star Alliance”
The “Star Alliance” that binds together United, Lufthansa Singapore Airlines and 13 other carriers provides a seamless, global travel solution to reach nearly 800 airports in 129 countries. These are multibillion-dollar companies that realized they needed partners to offer a friendlier, more valuable experience to their customers. The airlines have integrated capabilities and back office processes which enable global travel.

Imagine a software “Star Alliance,” with vendors pulled together around vertical domains or horizontal processes. The products would be held together with pre-integration, common GUIs and united go-to-market models.

Customers would get the full benefit of an alliance between smart companies, plus the extra benefit of innovation from individual vendor efforts. The alliance would enable vendors to band together in a sensible way and integrate common components. It would drive a revolution in buying behavior and possibly drive growth into the double digits again.

The part that is really significant, in my opinion, relates to the software “Star Alliance,” which is what I have been screaming in relation to the storage management software industry, but not using that specific analogy obviously. I chose the much geekier Storage Existentialism because that’s how I roll. That’s what it’s all about: a storage management software “Star Alliance.”

As frustrating as it is when people don’t get my story, it’s even more frustrating to find confirmation of my rants from some very smart people but at the same time discovering that VCs bet on the jockey, not the horse.

stor·age ex·is·ten·tial·ism :
n.

A philosophy that emphasizes the uniqueness and value of data in a hostile or indifferent plane, regards file systems alone as unexplainable, and stresses storage action based upon business value

I’ve been thinking quite a bit lately about viral applications like del.icio.us vs. the enterprise application market. Which direction does one take when confronted with some basic dilemmas of a startup? I would argue that the cost of creating and releasing an application focused on the enterprise market carries more risk though it may also have a commensurate reward. Even though the unprecedented access to skilled world labor markets presents a compelling opportunity to absorb less risk, the time required to come to market with an enterprise application is far greater than that of its big city cousin. Couple today’s requirements for seamless interoperability with broad platform support and you’ve got a QA nightmare on your hands when trying to develop applications that run in the data center. The Web 2.0 world of social networking and search can’t be bothered with such development and release roadblocks.

Enterprise apps can’t be in “beta forever” like the Web 2.0 applications of today. Technorati seems like they should be in beta though they are not. How long was Gmail in beta? Google is known for keeping something pre-GA for a long as they need. But, the engineering effort alone to get to beta phase for enterprise applications can be very expensive, especially for a cash strapped start-up. Then that leads me to think about the blokes from Skylook. These guys from Australia develop a plug-in to Outlook that integrates with the Skype API. These full-featured client plug-in applications relate far more to their Web 2.0 cousins than do those applications meant for utility and enterprise computing. Plug-ins require less engineering effort since they’re typically focused on being able to plug-into a limited set of applications and technologies. With Skylook, they only have one application to be worried about — Outlook. Their business is based upon Skype so they only have one API on that end to keep up with. But I don’t see plug-in developers making a huge impact on the world at large.

Pimpin’ ain’t easy. Especially when you’ve set your sights high and you’re making an attempt at big pimpin’. Alright, I’ve whined long enough about how much more effort and risk are required to sassify big businesses. I’m not really considering putting something out geared for the Web 2.0 world for a simple reason — I’m not in love with it. It doesn’t keep me up at night and consume my thoughts whilst sitting idle at a stop light. But data storage and information lifecycle management gets me randy. (That might be the geekiest thing I’ve thought all day.) I’m consumed by the thought of data having a beginning, middle, and an end.

But it’s not as simple as that. I’m further intrigued by the macro relationship of the lifecyle of information to the physical storage infrastructure on which it resides. It’s data storage existentialism. Humans have a variable lifecycle compared to other humans, but in the bigger picture, just accept for the moment that you can only discretely affect the course of this lifecycle. A human’s habitat and environment take into consideration the stage of the lifecycle and often times will evolve appropriately. That doesn’t really happen in today’s disconnected data center. The environment for data rarely evolves just because the information is no longer beneficial. Why? Because it’s not overly clear to anyone, especially those responsible for the environment, that the habitat can and should change accordingly.

There are quite a few companies out there trying to solve this problem, but their attempt is in such a monolithic fashion that it ends up competing with other better-suited remedies. It’s as if they’ve found what they consider to be the miracle drug that is the end-all-be-all, but you can’t compete with Tylenol on certain levels.

Or then on the other hand, you’ll find the companies that are under the mistaken notion that the entire problem of information lifecycle management falls squarely in their ballpark. (I hate sports analogies, so I’ll stop right there. You get the picture.) Mark-a-tecture in practice requires that companies focus their spin relative to what they do. I don’t blame them necessarily, but I don’t want to present myself as the end-all-be-all. I just want to be the necessary cog that’s required to make the information lifecyle wheel turn more efficiently.

It’s not just about efficiency today though. I want to build a better mouse trap that also keeps you from getting into trouble with the various government regulations that dictate how you manage your mice. More on the mice later…