To start with

  • Disclaimer
    The opinions expressed here are my personal opinions. Content published here is not read or approved in advance by EMC and does not necessarily reflect the views and opinions of EMC. This is my blog, it is not an EMC blog.

Pages of note

Music I'm listening to

Storagezilla’s blog

« October 2007 | Main | December 2007 »

November 2007

November 30, 2007

Big in 08.

SMB and consumer markets. You'll find digital ink by the digital barrel spilt all over the place about these two topics. So what's interesting about them? I mean people have been looking for the pot of gold at the end of the SMB rainbow for a while now. Some with less success than others.

Things to consider..

1: There's more money in the SMB market than just selling arrays.

Most SMBs don't have a room with racks and chillers so they're very much open to SaaS offerings where in exchange for a monthly fee, which can be forked over by anyone be they technical or not, everything is taken care at the end of a net connection somewhere else.

In this case the only requirement for being the firm's CIO is access to a credit card.

Of course SaaS isn't easy. It has a lot of moving parts which makes doing it well very hard and as such not everyone can play in that space but there's rich opportunities out there for those who can get it together.

2: The information explosion is unstructured.

In a world of consumer created content it sucks to be a database vendor.

Databases, like over paid DBAs, have long been at the heart of Enterprise IT but now we find that the fastest growing segment of the information business isn't anywhere near IT or the structured data stronghold. It's in the prosumer/consumer segment.

That's the real big deal here. People have been chasing SMBs for years but this could be the first time where companies who scaled down from the mainframe are taking an interest in seeing what you do with those photos and videos you've taken with your camera phone.

Well perhaps that's an over simplification. They're interested in the Exabytes of unstructured data which will be generated from that and other such devices over the next decade as well as all the meta-data which needs to go along with it.

Ever get the impression your at that point in time where everything starts changing? Well here we go again.

One more thing on a totally unrelated topic if you're a VMware person be sure to grab VMjuggler. It's "Stupid Virtual Machine Tricks" for people who stare at screens all day. :)

November 29, 2007

Up with iSCSI

Reading Tony A's latest I get the impression there's some two dimensional thinking in a three dimensional situation going on.

Yes EMC while certainly with a larger piece of the storage pie than any of it's competitors (And with it's massive iSCSI footprint. People tend to forget about that), could have used all that M&A money to buy up a lot of other array vendors over the years but didn't.

I've asked the "Why not?" question to Mount Olympus on at least two occasions and the answer relayed in a booming voice has always been the same. EMC is an Information Infrastructure company. It is not just a Storage Products company or just a Content Management company or just a Virtual Infrastructure company or a just Information Security company.

It's an Infrastructure company. It want's to win in every market it competes in but it's not a one trick pony.

What's left unsaid here is being just a storage platforms company back in the crash wasn't helpful. It nearly killed EMC stone dead when the bottom fell out of the array market and that should never, ever, be allowed to happen again.

Moving on regardless of what you might have heard EMC isn't down on iSCSI. IP Storage and IP SANs are a massive business for EMC hence all the dual and multi-protocol systems in the line up. While some folks get religious about the transport those who deal with customers on a day to day basis realise that ultimately you provide what the market requires.

I'm not in love with FC/FCoE/iSCSI/NAS or whatever. It's a case of choosing the right answer to the question in front of you.

Do I see any pure play iSCSI technology out there which I think EMC should have or should now stump up for? To be honest no. The UltraScale architecture has significant advantages over all of those other offerings. As well as being dual or multi-protocol it's easy to use, usually offers much more useable space and scales significantly higher than many of those iSCSI only systems.

As for suffering from not invented here syndrome I'll put that down to the fact that EMC has only made five acquisitions this year, down from nine last year. Yeah it's been a slow year when it came to making multi-billion dollar acquisitions with the mountains of new technologies and customers those bring I'll admit but I was under the impression that's what people wanted EMC to do in 07. 

So ultimately this comes back to the idea that EMC is somehow down on iSCSI. I don't think that's the case and it looks to me that EMC's competitors have been doing a nice job it making it appear like the company is the FC loving Lion in Winter who doesn't get what all those damn kids find "hip" about iSCSI.

While there might be a seasonal chill in the air outside your window it's not winter anywhere in the storage market yet my friends.

November 27, 2007

Fortress

Project Fortress? I mean really who makes this stuff up? I have no idea what you're talking abou... ;)   

Moving on I notice HDS have refreshed their rebranded BlueArc NAS gateways and dropped the price.

Field programmable gate arrays in a mid-range product anyone? So do I get non disruptive upgrades? I'm thinking not. I'm also thinking that even though they've dropped the price it's still a niche product. Yeah yeah sure sure it goes after the big ball of fun that is ONTAP GX (Which has it's own set of problems. I've had NetApp customers ask me about NetApp's plans for GX as their NetApp rep does some hand waving and quickly leaves the room anytime they ask about it) but regardless of how they price it it was still designed for a niche market.

This is HDS's fifth lash at NAS in six years. I'm kind of surprised that they didn't buy BlueArc the way they bought Archivas, but after six years of watching them not get IP Storage I should have seen it coming.

November 22, 2007

Marvel Digital Comics Unlimited: The Anti Long Tail

Last week to much fanfare Marvel announced Marvel Digital Comics Unlimited A subscription based service where comic book fans could trawl Marvel's back catalogue and read Spider-Man until they were ready to puke.

Having looked at the service since the servers became stable (Marvel.com ground to a halt seconds after the announcement. MySQL error messages popped up everywhere. Hey Donncha! Caching is the answer right?), I've had time to dig into the service.

It sucks.

Seriously.

If any industry stood to make a ton of cash from online distribution it was the insular comic book industry. Comic books are a backwater niche where 5% of the readers buy 80% of the books. That's a dead end market and a reason why television, movie and toy licensing are the areas which comic book publishers increasingly depend on to turn profit.

Superman is a profitable comic book franchise for DC but it's a hell of a lot more profitable for parent company Time Warner when you throw in the Smallville TV series, the Superman movies/Saturday morning cartoons and toy/apparel licensing of the "S" shield world wide. The same goes for Batman and Wonder Woman but the average age of comic book readers is increasing year over year not decreasing. Comic book publishers are failing to grow readership while the routes to market for their product have declined to such a point that a single distributor has a monopoly over the distribution of their printed product.

So how do you get comics into the hands of people not reading them? Go after the iTunes Music Store generation. Go after the digital download generation. Instead what we have in this version of MDCU is Web 1.0 thinking and good old fashioned protectionism of an antiquated direct sales model.

You subscribe to the service monthly and get access to their scanned archives there is no download option so you're grabbing pages from their servers with every page turn and reading them in their patent pending reader app. The problem I have with the reader app is that it's advanced functionality is a bit wacky and the resolution of the page scans is poor under magnification.

You'd think that a product which is digitized the moment the inking is done would allow for a high res image. You'd be right. Right now it looks that for various bandwidth and anti-piracy reasons they have gone a bit too lossy with the image compression for my tastes. But that's the point now isn't it? It's not supposed to look as good as it would had you gone out and bought a printed version.

They also appear to only be going for a back issue only strategy. Instead of this week's hot releases being available you'll be lucky if they've material which was on the shelves 12 months ago. Yet again a concession to dingy poorly stocked comic book stores.

Focusing on existing readers digital distribution would be a boon for a couple of reasons. Storage capacity (Hard Drives Vs Bookshelves) and the book you've heard good things about not being there when you make it to a store are just two of them.

Marvel publisher Dan Buckley has made rumblings that this is Marvel's answer to digital piracy of their products. That's only half true since both Marvel and DC have been issuing cease and desist letters to web sites hosting links to torrents and direct downloads on file sharing services such as RapidShare or Mega Upload for the past few weeks.

This is pointless since those sites just point to content already pirated shutting them down won't stop a single book from being pirated.

This is the music piracy war all over again and MDCU as it stands today is one of those subscription services which flamed out quickly back then. It doesn't offer any significant advantage over going to the trouble of tracking down this week's releases, downloading them and then buying something else at store later.

While some collectors will settle for a download of the monthly's they'll buy the much higher margin hardcover editions in dead tree format. Those are things you keep on book shelves not stacked in a box somewhere.

Since it's hard core readers who are doing all the pirating (This is a niche group with a specialized interest. If they're not downloading and sharing then they're scanning and uploading. It's a painstaking process to do correctly and not as simple as someone ripping CDs in their bedroom while running a file sharing program), you need to add features which will appeal to the hardcore.

Here's a thought. In an issue if someone (or the editors) reference something which happened in the past, lets say Bart Allen being beaten to death, I should be able to click on the speech bubble and go directly to the issue of The Flash where it happened. If it's an a-la-carte purchasing model they should offer to sell it to me as a download for a couple of bucks.

Do something like that and it's the long tail in action and it's something pirates can't offer as it's the publishers meta-data as a competitive weapon.

It's early days yet but just like all the other content folks I don't see the comic book industry getting the technology right. Content folks are the wrong people to come up with the distribution platform and that's why death, taxes and piracy are assured for as long as the planet keeps turning.

November 20, 2007

PR War. The Writers Guild hammer the Studios

With brain damaged reality television programming (To rip off a joke from 30 Rock, America's Next Top Pirate anyone?), the studios can wait out the writers strike but they're taking a massive PR hammering with every day that goes by.

The broadcast pipeline for scripted drama has not drained out yet so the bulk of the TV viewing public are unaware of the gathering storm. I wonder what'll happen when TV networks start entering re-run mode en-masse? Personally I'll be watching DVD box sets of TV shows I didn't watch the first time around.

Unfortunetly it looks like that leaves me with Melrose Place and JAG. Two of the shinest examples of crapola TV that I've ever seen and only one of them is from the guy who brought us Airwolf.

Note to self buy Airwolf box sets.

What's up with NeoScale?

JT added a comment a few days ago. The answer is something I've been sitting on for a number of weeks. I apologize for appearing to ignore the comment but it's not for me to repeat the contents of a private conversation I had weeks before hand when the info was non-public at the time.

Eyeofsauron

I heard that MTI had cratered shortly after it happened (Eye of Sauron folks) so it wasn't difficult to see who'd be sucked down with it, at this stage you should have heard that NeoScale went insolvent over night and have had to refocus. Just another startup story going badly wrong. It happens.

There's a ton of rumors floating around the place about who'll do what and when but I'm not willing to speculate on any of it as I don't have any extra insight (yet) as to who's speaking to whom about what.

What facts we have are that the encryption appliance business is "lumpy" (Dan Warmenhoven's word not mine) it hasn't turned into the license to print money that some people were expecting.

I'm on record of never being that much of a fan of the appliance approach due to port count and scalability issues those are geeky reasons to be down on something I'll admit but I've come to accept that it's a viable solution to some customers data at rest encryption problems. It however isn't a solution to all problems like some proponents of the approach would have you believe.

It's early days in the data encryption space so there's more than a few surprises, setbacks and reversals of fortune yet to occur.

November 19, 2007

Terry Tate: Office Linebacker

Someone hire this guy right now.

Amazon's $400 eReader

While I'm sure everyone out there is waiting for the next iPod I don't think this is it.

The original iPod was ridiculously overpriced and badly under spec'd when I bought it but in conjunction with iTunes it filled the synchronization and organization needs of my rapidly growing music collection. At the time it was a no brainer of a buy even if the price tag was silly all that being said I don't know what need Amazon's Kindle fills?

Books are pervasive and not a product undergoing large scale piracy. They also can require a substantial investment of time unlike pirated music, TV shows or movies which makes them a less desirable to the ADD piracy mass market.

Yes Kindle is nifty technology but it's no more a nifty technology than it was when I looked at Sony Reader a few years ago. The one big win I can see is that you could search your eBook collection, something I'd find useful since I've spent more than a few minutes leafing through dead tree format books looking for some fact, story or quotation I'm certain I've read but can't remember exactly.

None the less it won't be make your mind up time until I fly long haul or curl up in bed with one.

(That sounds filthier than I had intended it to be.)

November 16, 2007

VMware smack Oracle back

Just like Oracle's (To date failed) assault on RedHat was met with resistance VMware are meeting Oracle's VM announcement with steel as well.

To be clear, VMware does find processor paravirtualization to increase performance significantly on some workloads today, but the longer term performance delta when second generation hardware assist features are available is unclear. The performance difference may be reduced, eliminated, or expanded as enhancements to the paravirtualization interface may create new opportunities. It’s an open question.

As VMware sees it, the major problem with processor paravirtualization is the need for guest OS modification that makes it dependent on a specific hypervisor to run. The Xen interface, for example, implements deep paravirtualization with strong hypervisor dependency. The OS kernel is closely tied to structures in the hypervisor implementation. This creates an incompatibility as the XenLinux kernel can’t run on native hardware or other hypervisors, doubling the number of kernel distributions that have to be maintained. Additionally, it’s limited to newer, open source operating systems as the intrusive changes to the guest OS kernel require OS vendor support. Finally, the strong hypervisor dependency impedes the independent evolution of the kernel.

Now here's the thing, unlike Unbreakable Linux (Which just this week Larry has been promising will really take off over the next twelve months having been ignored by the market for the previous twelve months), where Oracle couldn't say they were supporting that and no other distro they have announced they will only support Oracle running virtualized on top of Oracle VM.

So they've just fragmented Xen even further and put themselves at odds with some of their largest customers.

I also think they've backed the wrong open source horse. The Linux Kernel Virtual Machine Project has much greater goodwill amongst the core Linux kernel team Torvalds committing it to the kernel after two months at a time when Xen had been waiting for more than two years.

Sure Oracle VM is good for Oracle's OnDemand SaaS offerings which are Oracle on Linux systems but that doesn't make it a fit for stand alone customer environments.

Just when you thought the Xen swimming pool was crowded enough in jumps Sun with xVM..and out jump some of their users. What some Solaris admins are saying is that they were hoping for an extension of Solaris Zones not another hypervisor flavor for them to check against the support matrix with whatever hardware software combo they're running with.

And remember if you run Oracle on it it's not supported.

November 14, 2007

What Maui isn't.

Anytime I see an EMC product codename pop up in the press it usually comes from one source. Joe Tucci.

But no matter how many fits he gives product management or how many sales people will scramble (sometimes to me) for more technical answers to customer questions JT is The Boss and he can say what he wants.

You've seen the start of the prosumer NAS initiative with LifeLine, Mamba you'll hear about soon enough and Hulk I'm just not going to talk about for fear my corpse will be dredged up from the river first thing tomorrow morning but I will touch on Maui just a bit. It's not Clustered NAS or anything like that. That's too narrow a focus.

Think oceans of data.

Think truly global information services.

Got it? Good. We'll talk about this again soon.