Storagezilla

Notes on a quarter

Couple of quick notes on EMCs quarter from the point of a stock holder.

V-Max sales took me by surprise.

It’s starting to look like the Symmetrix team aren’t happy maintaining the status quo.

Usually that’s a pretty static segment in the market, there’s a bit of customer substitution back and forth but rarely would any of the three major players in the high end face a rout.

28% year over year growth and posting a sequential increase in a quarter where the street had projected sales to be negative Vs Q4 shows there’s a shift going on somewhere in that market.

V-Max sales are trending above projections and the attach rate of FAST V1 is already kicking around at double digits in it’s first quarter of availability.

Symmetrix is looking to take more of everyone else's business.

Avamar and Data Domain still growing like crazy.

I’m going to be careful here as I know the exact numbers but the 100%+ year over year growth figure called out in the various press pieces is accurate. Relative to the other players in the backup market it would be interesting to see who’s growing their position relative to who’s shrinking. There’s only so much money in backup so if it’s landing in the coffers of one player it’s coming directly out of the pocket of another in either lost sales or cut to the bone discounts.

Discounts being good for the customer (and perhaps the sales person) but rotten for the vendor in the long term as it’s not like the product gets cheaper to develop or support just because the sticker price on the front of it has been marked down.

Now for a complete aside…

VMware’s financial model is looking different these days.

This is a comment on a completely different set of numbers but it looks to me that VMware’s revenue mix is moving towards a more subscription style model. A model which served RedHat very well in the recent down turn but who’s absence put the hammer on Microsoft’s server software revenue during the financial doldrums.

I suppose we’ll know more about this during the next major VMware product release but even that looks to be changing with them looking to deliver a product instead of a sever component now a management component then and middleware whenever.

I’m not from Corporate but I am here to help.

EMC World used to be so easy. When it was Wizards way back when it was enough to get people under one roof and whatever happened, happened. Then with a bit of spit polish it became EMC World, all the Execs would come streaming out of the Puzzle Palace down in 176 South Street, there would be an announcement or two and we’d all go home.

I know exactly what happened this year but there’s still too much going on. It doesn’t matter if you’re out in the furthest reaches (Me) or you’re “from corporate and you’re here to help” (Them) if the people you’re speaking to aren’t involved in EMC World this year they’re probably directly working with someone who is.

Biggest EMC World ever.

Malware is an Enterprise problem

I was sitting in a looonnngggg multi presenter meeting eyes somewhat glazed over when I was suddenly snapped back to reality by one fact.

88% of the US Fortune 500 domains have been accessed by systems infected by the Zeus Trojan.

So the bulk of the Fortune 500 has been touched by or infected with the Zeus Botnet.

The bigger the organization the larger the attack surface the greater the chance employees are accessing corporate systems using infected machines externally or internally.

You can get infected with malware from advertisements on things as benign as social networking sites, you don't even need to be using a social networking site, personal and work email are also distribution vectors so it's not like employees have to be surfing for pornography to track something nasty into the company on the soles of their digital feet.

Do anti-virus packages solve the problem? Well they can help, but they're not the full answer. I've accessed the corporate network from my home computers. They're all running up to date anti-virus and anti-malware protection but it's already been shown that the Zeus Trojan is only detected by major packages less than a quarter of the time as criminals modify and morph it to evade detection. So you won't know what's lurking inside your organization until you sit down and start gathering intelligence.

Malware is like an STI, you might not notice it until you have a flare up, if they're not treated they can spread and people don't run around the place advertising the fact they're infected.

And neither is healthy or hygienic.

EMC has a ton of white hats and mathematicians working  on taking apart and analyzing Phising kits, their attacks, servers and networks as well as Trojans and the giant zombie Botnets they're slaved to at RSA FraudAction Lab. Now they're turning that research into a service which can be used to analyze infrastructure and dig out what information has been comprised by an attack or infection inside your organization.

You can find out more here.

The Great Products

“From the splash screen to the screws I want to build great products.”

I’ve been asked numerous times before what my career plan is and my answer has always been “I don’t know”, that’s been a cop out for my real answer as given above. My answer is somewhat quirky as it doesn’t involve ambitions of a title, hordes of minions and a corner office.

If I actually had business cards printed, which I never do, in lieu of a title I’d put “From the splash screen to the screws I want to build great products.” It’s why I focus on work stuff beyond my day to day remit and send a blizzard of emails at all hours of the day and night to people I’ve never physically met and probably shouldn’t be mailing in the first place.

So become a Product Manager!

No thanks.

Product management involves pragmatism and compromise, you can’t march into a business review and say “I plan on tanking revenue, losing customers and handing a perceived lead to the competition for 18 months to two years while I finish this masterpiece I have the development team working on.” Sometimes you can do that, most times you can’t. You won’t last very long in the role unless every product is a world beating product at the end of that process.

In general how you get along is to align your goals with the company’s goals while understanding that every publically traded company’s primary goal is to maximise shareholder return. It’s not to make great art or the best shampoo or the finest products it’s to ensure the owners get their money’s worth from their investment.

The company might do that by making great art or the best shampoo or the finest products but somewhere and at some time if you’re a Product Manager you’re going to have to slowly swallow a pint glass full of Castor oil or do something which gives you that sensation just to get along.

My way of going from place to place offering solicited and unsolicited input and ideas suits me better. Most certainly I’m not happy when things don’t get done for various reasons or things take longer than I’d hoped. In general everything takes longer than you’d hope, for example I didn’t know that half the amount of time spent bringing something to market is spent on QA alone, but you do what you can where and when you can.

Now all I have to do is balance out my non-work related interests and I’ll be doing well all round.

You can spend too much of your time and your life at this stuff.

NetApp performance testing engineer tells it like it is

The other sales team told my client that they would actually get more IOPS from the NetApp solution than ours because of their Performance Accelerator Module II (PAM II).  They configured the 65% SATA and 35% FC solution with the 512GB PAM II card.  If you use the scale mentioned above and take into consideration the drive spindle count and drive type, we calculated that their IOPS would have been approximately 37,000, without the effect of the PAM II card.  However, the NetApp team told my customer that due to the large amount of cache on the PAM II, and the nature of the workload, obtaining 65000 IOPS or more would not be an issue.

They also told my customer that due to NetApp’s deduplication ratios, they could cut back on the amount of storage that had to be purchased.  My customer needed 147TB of storage for current needs and growth.  The NetApp solution only provided 102TB with the majority of it SATA storage.  The justification was that due to NetApp’s ability to deduplicate everything on the box, including production data such as Oracle, SQL and Exchange; my customer would save lots of money because fewer disks needed to be purchased.

The NetApp solution seemed too good to be true.  It seems that the NetApp solution was equivalent to our solution but with less storage, less spindles, and less cost.  With that, I set out to get more educated on the NetApp solution.  Instead of reaching for the competitive section of EMC Powerlink (EMC’s information portal for partners, customers and employees), I placed myself squarely in the customer’s shoes and went right for the NetApp documentation library.

via blog.dpweiss.com

Sounds pretty much like it came straight out of the NetApp song sheet right?

But what follows later in that post is a great example of what happens when you do your own research.

Bycast customers, we’ve got you covered

NetApp’s acquisition of Bycast puts HP and IBM in an interesting bind. We know of about 100 customers using Bycast running on HP (HP MAS) and IBM (IBM GMAS) servers. Maybe the IBM-NetApp OEM arrangement will resolve IBMs stance on this but I can see HP giving NetApp-Bycast the stink eye already.

Never mind. EMC has been providing object storage solutions to thousands of customers for more than a decade.

Having seen Bycast in deployment myself the first thing I noticed was the triple management layer of Universal Access, Grid Management and Archive Management. It’s hardly a turnkey solution to deploy and manage it also struck me as odd that the product be called a grid when in reality it’s just one application running on a cluster and doesn’t realise any of the benefits of true Grid computing.

But now I'm just nitpicking at marketing spin.

Regardless history has shown NetApp will promise it’ll be fully integrated into ONTAP about 18 months from now which means they’ll do a press release saying it’s done in 2016.

But they’ll not deliver the code until the point release after that.

Alas poor Spinnaker..