This evening I attended a company session hosted by NetApp in their office in Schiphol-Rijk, the Netherlands. It was quit a trip, due to all sorts of reasons I decided to go by train. At first I was somewhat sceptic about going by train, but by the time I was half way there I felt it was quite nice, I even closed my eyes for a couple of minutes.
Looking out the window on the motorway lying just next to the railroad track I was even more pleased with my decision taking the train. I saw a major traffic jam near Amsterdam which I probably had to deal with when I travelled by car.
The traveling time by train was significally longer (when the railyway track was build there was no motorway where there is now, back then it was all water.) almost two and a half hours and that was one of my reasons not to take the train. But when a collegue offered a ride half way the trip home I was eventually pursuaded to take the train.
If I travelled by car you probably did not get to see this message due to lack of time, I’m actually writing this offline in the train.
Ok, enough travel talk.
Back to NetApp, they are located at the businesspark at Schiphol-Rijk which is next to Schiphol International Airport. Nice to know a lot of major US-based technology firms also have an office here e.g. Microsoft, Juniper etc. Because it is so close to the airport it is an attractive location for firms that do have employees who travel a lot between different locations by plane.
The evening started with a dinner in the restaurant which appeared to be the session room also (my compliments to the catering it was excellent!). The building is relatively new and well equipped as far I could tell of course (only been at the restaurant). And believe me dinner came at the right time for me because after taken the train… – now that was a slip of the tongue.
After being welcomed by our collegue Marcel and our host Bram of NetApp it was time to get rolling. Amzar kicked off with a outline of the proof of concept (POC) he did at a large international customer with NetApp storage. Goal for this POC was to determine if VI-3 at remote sites (10-25 servers) could rely on ISCSI and NFS storage compared to the FC connected SAN storage used at the large sites. The interesting conclusion of the POC was the advice of using NFS for storage at remote sites. The nice thing in that situation is that they already use NetApp solutions at remote sites. As NetApp explained later on the boxes can present almost every kind of storage to a client from NFS,ISCSI FC and even CIFS for example homedirs. The desicion to advise NFS was based on flexibility (re-use of existing hardware and simple expansion of NFS volumes), cost effects and simple file based restore. Performance was not an issue.
The next presentation by Mark from NetApp covered the NetApp products and solutions as there are differentiated in Store, Manage, Protect, Retain and Secure. I’m not going to deliberate on that for more information I would like to refer to the channel guide which outlines NetApp Products and Solutions.
One thing which I think is important to adress is the uniform managebility across all storage products levels. There is no difference in management of entry level systems compared to mid and high-end level, so no training issues when you decide to upgrade to higher (or lowel) level systems. The determining factor choosing the system level most likely will be the amount of IO’s per second you need.
Also very important is the deduplication of data which is now free of charge (announced last week) and delivered out of the box, this can reduce your storage space up to 50%. It works at block level and searches for duplicate blocksets and uses pointers to adress the data, also usable for homedirs (e.g. CIFS).
Not covered in the session but to give you an idea about the power of deduplication take a look at the video I found on youtube: a nice demo of deduplication in a virtual infrastructure in combination with VDI:
Also announced is Snapmanager for Virtual Infrastructure; quote from same article: “which lets users take snapshots of VMware virtual machines. “It’s fully integrated with VMware and it allows you to use NetApp snapshot and restore capabilities in a VMware environment,” says Rogers. “[Before] there wasn’t a simple, non-disruptive way to snapshot a VM.” “. Snapmanger will be released for the public in May 2008.
The last presentation of the evening was from Jan also from NetApp taking a deep dive in technology. Interesting was the comparison of VMFS (on FC) and NFS showing lower (better) latencies with NFS and higher throughput (400 mb/s) compared to vmfs (180mb/s). Also nice was their Raid configuration called RAID DP (default 14 + 2 ) which is 200 times as reliable as 2 times RAID 5 (7 + 1disk) for RAIDgroups.
Concluding: A very nice Session at a nice setting and lots of information.
At April 3 2008 NetApp organizes an annual one day (free) event called “Innovation 2008” at Van Nelle Ontwerpfabriek in Rotterdam in the Netherlands. For more info and registration look at www.netapp-innovation.nl .