Hi, I’m super excited to write these series of blog posts about the new XtremIO generation we have just announced, XtremIO X2. I won’t cover every aspect of the new […]
I’m super excited to write these series of blog posts about the new XtremIO generation we have just announced, XtremIO X2. I won’t cover every aspect of the new product but rather try to give you my CTO’ish overview, if there is one word I can describe it, it will be “Maturity”, This project was a very ambitious one, we already had a first hit record (XtremIO X1), an huge install base, the largest in the world ( https://www.emc.com/collateral/analyst-reports/emc-xtremio-ranks-1-research-proves-it.pdf ) and so, while we had to take care of our beloved install base, we also started to develop what would eventually be called “X2”
Below, you can see a part of our huge lab in Durham. Walking through the lab is giving you the “Matrix” kind of feeling. The site in Durham is only one lab out of many we have and that is an important thing that we could only achieve with the budget of EMC and later on, the DellEMC R&D Budget.
When we started building XtremIO in 2009 (god, how time goes by!), we could choose between the easy way and the hard way
Easy meant developing a Scale-UP architecture, a typical dual controller architecture that can only scale with capacity which meant that if you need more performance than what the array storage controllers can provide, you need to buy a new storage frame that won’t share the capacity of the performance with the array you already have, the benefits for this type of architecture are in favor for the R&D team only, it’s a much shorted R&D development, however, as a customer that will require the performance one day? Not so much.
Hard meant developing a Scale-Out architecture, one where you have more than a dual controller’s architecture which can squeeze every bit of performance from all the available storage controllers in the system, we of course chose the hard way. I wrote many posts on the topic, including, XtremIO 3.0–Why Scale Out matters? ,
A Tale of Two Architectures — Engineering for the 99.999% versus the 0.001%
and if it wasn’t obvious, we chose the hard ward for us which meant the RIGHT way for you the customer, no one is dare asking “why do you need 1M IOPS or why do you need sub ms latency” anymore.
Product development is a never-ending cycle
While XtremIO boomed in the market place, we knew that there are areas to improve, mainly around density, additional data services and even better stability as the array is hosting the most mission critical applications out there.
Years of R&D development were invested in this platform as we will cover the highlights in the upcoming posts, below you can see my solutions pod hosting single X1 arrays, single X2 arrays and Dell FX2 blades.
The People or, the MOST important thing you company has
Such an huge task like X2 is an astounding project that can only be achieved with dedicated engineers, countless meetings, countless architectural meetings, lots of code lines, its like a family with many parents, I can’t tell you how many times I emailed people in the middle of the night after finding some bugs during my testing and the fix was waiting for me in the morning, guys and girls, I salute you!
Here’s the index for the following X2 Posts: