What were the storage trends of 2009 that you think will continue to expand in 2010?
Deduplication is a big one. There is a lot of interest in dedupe. It is going to be increasing as data is increasingly dramatically. This also [means] archiving - where you put your data - is going to increase.
Tiered storage is another area. Again this comes from the economic issue [with customers] having to optimise how they distributed data and storage. They can't keep it all on [primary], you have to move it around to cheaper storage.
It is going to continue to be big but I was at a customer conference in Europe last year and I brought up this question of tiering. I said "I am sure you guys have looked at tiering" and they said "yes, we have but it is too much of a headache as now you are leaving it to me to decide when to move data where and I don't want to do that."
I said "why don't you leave it to the end user?" Well the end user doesn't care. As far as the end user is concerned, [they want you to] keep their data on the fastest possible media and tier everyone else's. So no-one is going to move the data themselves.
Then the administrator has to move the data. You can use policies like after 30 days move it here etc but there is data that is more valuable and retains its need for urgency no matter how old it is So I think software that allows for that movement will be important.
So does Dell plan to get involved more in this area?
Dell is taking all the customer input and coming up with solutions. Whether the solutions are entirely made by Dell, partnered or whether we buy a company to get that solution it is up in the air [but] for all these areas we are going to definitely add products for our customers."
Where does virtualisation fit in when it comes to storage? Is it still holding back?
With server virtualisation, the idea behind [it] is to take one server and chop it up into lots of little bits [yet] storage virtualisation wants to take lots of little bits of storage and make it look like one big bit. It is the total opposite execution. You are still abstracting from the physical but you are doing it differently.
Most companies have implemented or are implementing server virtualisation for the benefits of increasing utilisation. If you do that then you actually run into problems with your storage as it wasn't designed for it. So that interlink has to be optimised.
If you look at storage virtualisation, the idea is to take lots of separate storage boxes and make it look like one big box. Why? Because it is easier to manage one box than lots of little ones and also because usually when you start filling up a file system and you fill up one box, what do you do with the rest? You can't break it and put it in another box. It is too hard to access those two things. So you want it to look like one big box so you can break it up. This is what we do with Equalogic.
Has there been a big take-up in Equalogic technology since you bought it two years ago?
It has done really well. I think when we bought it we had in the range of 2,000 customers or so and now we have 12,000. We had very few customers in Europe - just a handful - and now we have 4,000 so it has been very successful for us.
I think that it is doing well with customers who bought server virtualisation, it is doing well with customers who want to save money and it is doing well with customers who don't have a lot of IT skills.
