How is Nimble compared to other storage solution in terms of Performance and Administration?
Excellent and dead easy.
It's all relative at the end of the day. Nimble delivers extreme simplicity in storage (it's designed for the IT generalist) and has created a unique, patented file system which takes the performance and capacity scaling challenges that traditional vendors suffer from (ie Dell, HP, IBM, EMC, Netapp etc). And don't forget the Infosight data analytics which is far ahead of anything else any other vendor can provide from a support and proactive analysis standpoint (IMO, of course!)
There are numerous reviews of the tech on the net - Reddit is a good place to look. Check out this thread: Anyone here using Nimble storage? : sysadmin, here's a quote:
"I've installed many nimble devices in the last 18 months, and they are by far the easiest and quickest to rack, and configure!!!
I've done an installation for customers in less than an hour, and they think I'm joking when I say I done.
The usability of the interface is clear and everything is where it should be.
I particularly like the performance stats you get on the management website out of the box, unlike HP and EMC devices where another piece of software at a cost is needed,
Then there is infosight, the web based performance and incident monitoring, which is included in the purchase of your device. This allows the Nimble to call home and report faults and when the capacity is full so you can order the upgrades. Yes other vendors such as Dell offer this feature, but its a pain to setup with SANHQ etc etc. Where as this is all built into the Nimble.
Finally the last part I love about the Nimble, its modularity is perfect for small customers. For example, if you buy the CS300 module, and you out grow the performance, rather than other vendors where you buy the next module up, find the rack space for it, move your data over etc etc.
You simply replace the components you need to upgrade, whilst the system is live, yes no down time, this includes, drives, ssd cache, controllers.
So tbh I'd recommend a Nimble first before any other storage vendor, based on most peoples needs.
There is a whole other subject of price I could go into as well, but I'll let you be the judge of that against other vendors entry level devices, and comparing the bang per buck!"
I tested TinTri, Nimble and Tegile last Spring/Summer. At the time we were using a variety of storage, mostly EqualLogic and EMC VNX/NS arrays. I would have been happy with all three. They are all better for us than EMC/EQL by a long shot. We are a smaller IT shop, no dedicated storage people, etc., so while EMC management isn't hard, you do need to use it regularly or the complexity will thwart you.
Take what I say with a grain of salt. This is just my opinion in our particular situation, and everyone has different needs.
Ease of use: TinTri > Nimble > Tegile - TinTri is very, very easy. Plug it in, configure three NFS parameters on each ESXi host, reboot. Done (OK, you could get fancy with network LACP). Nimble is close, requiring an ESXi plugin, plus more cumbersome vSwitch configuration. Nothing hard, and on my second host it was done in about 15 minutes. Tegile was a challenge, but still easier than our VNX. Their current release should be much easier, according to our SE. Operationally the ranking still applies. No thought required with TinTri. Some for Nimble, but it is all fairly straight forward. Tegile is also easy to use day to day, once you get it setup.
Flexibility: Tegile > Nimble > TinTri - TinTri is a one trick pony. Well, two now. Tegile can solve World Hunger and Peace in one array.
Performance: They are all similar, and way faster than our VNX in most situations. The two caching arrays (Nimble and Tegile) have very different performance characteristics than TinTri. For raw performance it is about sizing your array for all three of them. I'd say the TinTri experience is more consistent with their awesome QoS engine. It really works well. All three of these are hybrid, so if you find the right parameters and have to hit disk, these will fall over. I was able to make all three fall over with very specific synthetic workloads, but I don't think I will ever be able to do that in the real world. When you do hit their performance limitations, then you are dealing with spinning disk. At that point, more is better. Our VNX has 100+ disks, so you when you exhaust it's anemic cache, no big deal since you have so many spindles. It is a graceful degradation. These hybrid arrays hit their performance wall more sharply. It's the same as always, know your workload and buy accordingly. It helps to have a good VAR to help you. I will never go back to a standard array.
While we picked Nimble in the end, I would have gladly chosen the other two with a different set of needs.
What else can I tell you?
Doing NFS over 1gb with VMware to try and mitigate the lack of multipathing isn't fun, and easy to get wrong.
I enjoyed the challenge getting it right with our NetApp, with help from various guides from the NFS guru Chris Wahl (you need a lot of info at the VMware, Cisco (or other) and the storage array side to pull it together), but at the end of the day it's still not as good as MPIO with iSCSI on a 1gb network.
We did have professsional services to install our NetApp, but the engineer was skilled in NetApp, not networking or VMware.
I can tell you that Nimble is almost as easy as EqualLogic, the original leader in easy to use SAN for the SMB market. Performance is way better than EqualLogic, which should no longer be purchased for new deployments. I found Tegile to be very difficult.
I've installed several nimble setups in the last months and the default installation en setup is strait forward is you follow the standard procedures.
The web interface is great the moment you see is you will find your way trough it like you already have worked for years with the product.
you don't need any special software for managing the array you only need a browser and flash and you can begin.
we have migrated from a netapp 3240( with flash cache) to a nimble 540 and went back from half a rack to 6U ( controller and disk shelff ) with better performance ( did a lot of IO testing because couldn't believe it when i heard the CASL architeture the first time).
We have tested other vendors but because of the stability of the POC unit whe have tested on.
next to that every environment is different I don't know how many vm's you have ore how many data you have to store.
you will need the right sizing to begin but keep in mind that you can almost always do an uninterrupted hardware upgrade off you system.
Retrieving data ...