Server-Based Applications for SSDs

[jpishgar]

Hey all!

Have had a few folks asking about this, so wanted to start this thread up as a possible resource for any IT professional-level users of ours out there who have bumped into this. In our IT Pro area, as well as here in Storage, we see a lot of talk about applications suited to SSDs. A number of the discussions have revolved around server-based applications. Which should you run on which type of storage media? That's always been a toughie and IT Managers the world over have been struggling to codify a setup that makes sense for their servers. Naturally, you want to make sure you've got the right applications running on the right hardware - especially in an enterprise environment. So we want your input on this.

So I put this question to our resident experts:
In a server environment, there are a lot of applications more strongly suited to SSDs - but specifically, what apps do you find actually realize the most benefit?

Chime in below with your thoughts, and since this is a spotlight thread, please note that contributors to the discussion may be highlighted in a feature piece later on (particularly if your comment is a good one!).

-JP

 

[popatim]

my #1 - database servers.
#2 - cache drives.

 

[Saga Lout]

This is more of a negative than a suggestion to avoid Windows 8 until they stop their Metro apps heading straight for the SSD with no option to install on a larger hard disk. This might matter less when larger SSDs become less expensive but at the moment, the clutter factor those apps cause on the SSD is immense.

 

[rubix_1011]

One issue that is enterprise-related is the cost per unit and then the time/energy/cost to implement. If you consider that most major enterprises run very large SANs or variants of such, the price and performance per GB and TB is fairly low $-wise and fairly high in I/O vs. your typical local blade or server box RAID as they utilize multi LUN storage over 10 gig fiber (example). If you were to implement SSDs in this space, you might see some Read/Write performance, but this all depends on how your SAN is setup. I'm not a storage expert, but I do know that when it comes to large system storage that isn't hyper-critical, platter storage will be the choice for more capacity in GB/$ than SSDs and likely will be until they level out at similar capacity/price points. Where SSDs really shine is when they are a direct 1:1 replacement for small RAID or single drives.

 

[jpishgar]

Where SSDs really shine is when they are a direct 1:1 replacement for small RAID or single drives.

If you were going to start from scratch though, and had to load out a NOC from start to finish, would you go SSD or HD? Or would you mix it up?

I'd almost think that an IT Manager taking a good hard look at what it would cost to migrate later to SSDs might go the tiered route to save on the headache later - and that the cost would be justified in the mid-to-long term. This would depend on the applications you are running though. An ecommerce-heavy server need would differ pretty strongly from a documents heavy backup/archival set of requirements, no?

 

[rubix_1011]

Yes, and I would imagine you would get the best results from a mixture of dedicated SSD storage assignments and HDD storage assignments independent of one another unless money is absolutely no object and you can afford a datacenter full of solid state drives. With an adequate storage solution and setup, many normal operations will always continue just fine with larger capacity, slower HDD's, but including arrays of SSD's can offer vital speed improvements on mission-critical nodes and systems requiring high I/O.

It always comes down to cost, and the simple fact that SSD's are still more expensive $/GB vs. HDD's is a huge, primary factor for anyone seeking budget approval, as capacity often wins the enterprise storage debate due to RAID and SAN technologies and distributed data access. However, as SSD's start scaling better in cost and capacity vs. platter drives, their offset can also additionally be debated against the proportionate savings in power consumption. Again, just my thoughts and no expert, first-hand knowledge...as these are also the same reasons I'm not on the SAN/Storage team.

 

[williamvw]

This is more of a negative than a suggestion to avoid Windows 8 until they stop their Metro apps heading straight for the SSD with no option to install on a larger hard disk. This might matter less when larger SSDs become less expensive but at the moment, the clutter factor those apps cause on the SSD is immense.

Hi, Saga. I may be the last guy to get the memo, but I hadn't heard of this before. Any clue why Microsoft might show this bias for Metro apps?

 

[Saga Lout]

No idea at all, I'm afraid. Hopefully W9 will fix it but I'm unsure whether anything as basic as hard disk speed is one of their concerns. Whizzy touch screen apps - that's another matter. My comment is almost irrelevant to the IT Pro's we're dealing with here but it is an issue to smaller businesses and domestic users.

 

[jpishgar]

Ecommerce is a given. Database is a given.

What about cloud storage for cycle-heavy render programmers? Or productivity items? Exchange server for certain, probably at anything 2,000+ users and above the benefits are pretty clear.

Hoping to get feedback here and be able to draw a distinction for folks trying to make a choice their company is going to have live with for a while when it comes to app selection. What to run on SSD, and what to run on SSD?

-JP

 

[Santanu_Saha]

Many years ago, I helped bring a product to market to provide third party caching with RAM SSD. I believed in the product and was able to get many others to believe in the product. What I was not able to do was to get many people to buy the product. As I look at solutions on the market, I can see that companies trying to sell third party caching solutions are encountering the same obstacles and are fixing or working around the problems.
Here are some problems I have experienced with third party caching solutions:
RAID systems increasingly offered reasonable cache levels narrowing down the field of customers that need write caching. At the time we offered this RAM write cache, we thought that Xiotech customers were the perfect target as they did not believe in write caching at the time. Fact is, the combined solution worked out pretty well but was only useful until Xiotech realized that offering their own write cache could solve most customer problems.

Given the challenges with write caching it seems to me that most companies today are focused on read caching. Read caching solutions have a long history. Gear 6 was one of the first to take the space seriously and had some limited success with environments such as oil & gas HPC and rendering.