Tag Archives: SSD

How to design a Scale Out NAS Architecture

Scale-out architecture and why it is important when architecting a storage solution.

I had an interesting discussion with an architectural firm the other day.  Most of the discussion was around scaling for the future.  In our discussion we talked about the linear scalability of the ISE technology and he pointed out that while that made a ton of sense for his block-access requirements he was a little concerned around the unstructured data, as well as some plans utilizing NFS for some of his server and desktop virtualization needs.  The last thing he wanted to worry about was changing his architecture in 12 to 24 months due to growth or technology changes.  So we started working on architecting a solution utilizing our new “scale-out” ISE-NAS solution.

You’ve probably heard a lot about scale-out type architectures. 3PAR sort of led the way with their ability to scale out (at least to eight) their storage controllers to their fixed-backend backplane-attached disk drives and it offers up a pretty unique solution (at least in a block storage architecture).  3PARs problem is they don’t really have an answer for the same scalability around unstructured data (NAS).  Don’t get me wrong, they list 5 NAS companies on their website, 1 is out of business and the other 4 have either been acquired by their competitors or is a straight up competitor.  This scale out architecture seems to have caught on in the emerging NAS Gateway devices like Symantec FileStore and Isilon.  Clearly both FileStore and Isilon are very different on the scale-out architecture.  More below.

 

So first things first, let’s describe what a “scale-out” architecture means, at least to me that is.  When architecting solutions, it’s always important to put a solution together that can grow with the business.  In other words, they know what they need today, and they have an idea what they might need in 12 months, but 24 – 48 is a complete crap shoot.  They could be 5X the size, or just 2X the size but the architecture needs to be in place to support either direction. What is sometimes not discussed is what happens when you run out of either front-side processing power, backend IOPS or usable capacity?  Most storage solutions give you 1 to 2(ea) clustered controllers, and a fixed number of disk-drives they can scale to dependent on the specific controller you purchase.  From a front-end NAS solution most of them only scale to 2 nodes as well.  If you need more processing power,  more backend IOPS or capacity, you buy a second storage solution or you spend money to upgrade storage controllers that are not even remotely close to being amortized off the CFO’s books.  If you look at the drawing above, you can clearly see what scale-out architecture should look like.  You need more front-side processing, no problem.  You need more backend IOPS or Capacity, no problem.  They scale independently of each other.  There is no longer the case of “You love your first <insert storage/NAS solution of choice> and you hate your third, fourth etc etc. Isilon is probably a great example of that.  They tout their “scale-out” architecture but it clearly has some caveats.  For example, If you need more processing power, buy another Isilon, you need more capacity buy another Isilon, you need more backside IOPS…well you get the idea :)  It’s not a very efficient “scale-out” architecture.  It’s closer to a Scale up !!

Let’s also not loose site on the fact that this is a solution that will need to be in place for about 4 to 5 years, or the amount of time in which your company will amortize it.   The last thing you want to have to worry about is a controller upgrade, or net-new purchase because you didn’t size correctly or you under/over guessed your growth or even worse, years 4 and 5 hardware maintenance.  This is especially true if the vendor “end of life’d” their product before it was written off the books !!!  Cha-CHING.

 

So this company I was working with fluctuates with employees depending on what jobs they are working on.  It could go from 50 people to 500 people in a moment’s notice and while they would LOVE to size for 500, most of the time they were around 50 to 100.  So as I mentioned above, we started architecting a solution that incorporated our ISE-NAS solution based on Symantec’s FileStore product. When coupled with our Emprise 5000 (ISE) gives them the perfect scale-out solution.  They can start with 2-nodes and grow to 16 by simply adding NAS engines (x86) to the front end.  If they need more capacity, or backend IOPS, we can scale any direction independent of the rest of the solution.  Coupled with our predictable performance we gave them the ultimate ability to size for today, and know exactly what they can scale to in the future.

In the world of “Unified Storage”, cloud computing and 3 to 5 year project plans, its important to consider architecture when designing a solution to plan for the future.  Scale-Out architecture just makes a lot of sense.  BUT – do your homework.  Just because they say “scale-out” doesn’t really mean they are the same.  Dual-Clustered controllers – or even eight-way – will eventually become the bottle neck and the last thing you want to worry about is having to do a wholesale swap-out/upgrade of your controller nodes to remove the bottleneck or worse, have to buy a second (or third) storage solution to manage!!

@StorageTexan

10,000 Exchange Users in 3U of space

 

This is a pretty cool video done by the Technical Marketing team at Xiotech.  10,000 Exchange users in 3U of space!!   No fancy/expensive SSD needed for this !!

In summary:

250 VDI instances or

10,000 Exchange Users or

750 DVD Quality Video’s or

25,000 MP3′s

In 3U of space.  Now that’s WICKED FAST !!!

 As they say in Minny – “that’s not to shabby !!”

By the way, if by chance 10,000 is just not enough users for you.  Don’t worry, add a second ISE and DOUBLE IT TO 20,000.  Need 30,000, then add a THIRD ISE.  100,000 users in 10 ISE or 30U of RackSpace.  Sniff Sniff….I love it !!!!!!!!!!!!

By the way – Check out what others are doing:

Pillar Data = 8,500 Exchange Users with 24GB of Cache !!!  I should say, our ISE comes with 1GB.  It’s not the size that counts, it’s HOW YOU USE IT !! :)

One Pillar Axiom 600 with one FC Slammer
24GB of cache <—-  WOW !!!!!
4 SSD Bricks for databases. Each with:
Two dedicated RAID controllers
13 50GB SSDs <— I’m going to guess that these aren’t very cheap.
2 SATA bricks for Logs. Each with:
13 500GB 7,200 RPM SATA disk drives

Hitachi AMS 2300 = 10,800 users – 400+ Pages PLUS !!! <– I have to say it again, WOW 400+ pages on this bad boy !!!

240 300GB 15K RPM SAS disks, <— Ahh ya  – TONZ of spindles !!!  We had 20(ea) 3.5″ Drives to do our testing. 
16GB of cache and
8(ea) 4Gb/s Fibre Channel paths was used for these tests.
Testing used 8 Sun Fire 4600 M2 servers with 32GB of RAM,
four dual-core AMD Opteron CPUs,
8(ea) Emulex 4Gbit/s Fibre Channel adapters and
Windows Server 2003 R2 Enterprise x64 with Service Pack 2.

The old pain in the DAS

The old pain in the DAS !! 

Is it me, or are others in the field seeing more and more companies being pushed to look at DAS solutions for their application environments?  It strikes me as pretty interesting.  I’m sure mostly it’s positioned around reducing the overall cost of the solution, but I also think it has a lot to do with assuring predictability around performance that you can expect when not having to fight for IOPS among other applications.  In fact, Devin Ganger did a great blog post around this very subject in regards to Exchange 2010.  It was a pretty cool read. I left a comment on his site, it pretty much matches (some may call it plagiarizing myself :) ) my discussion here but I’ve had some time to expand a little more.  As such, here are my  thoughts on my own site :)

Let’s take a look back 10 years or so.  DAS solutions at the time signified low cost, low-to moderate performance and basic RAID-enabled reliability, administration time overhead, downtime for any required modification, as well as an inability to scale.  Pick any 5 of those 6 and I think most of us can agree on it.  Not to mention the DAS solution was missing key features that the applications vendors didn’t include like full block copies, replication, deduplication, etc.  Back then we spent a lot of time educating people on the benefits of SAN over DAS.  We touted the ability to put all your storage eggs in one highly reliable, networked, very redundant, easy-to-replicate-to-a-DR site-solution, which could be shared amongst many servers, to gain utilization efficiency.  We talked about cool features like “Boot from SAN” as well as full block Snapshots, replicating your data around the world and the only real way to do that was via a Storage array with those features. 

Fast forward to today and the storage array controllers are not only doing RAID and Cache protection (which is super important), they also doing thin provisioning, CDP, replication, dedupe (in some cases), snapshots (full copy and COW or ROW), multi-tier scheduled migration, CIFS, NFS, FC, iSCSI, FCoE, etc etc.  It’s getting to the point that performance predictability is pretty much going away not to mention, it takes a spreadsheet to understand the licensing that goes along with these features.  Reliability of the code, and mixing of different technologies (1GB, 2GB, 4GB, FC Drive bays, SAS connections, SATA connections, JBODs, SBODs, loops) as well as all the various “plumbing” connectivity options most arrays offer today is not making it any more stable.  2TB drive rebuild times are a great example of adding even more for controllers to handle.  Not to mention, the fundamental building block of a Storage array is data protection.  Rob Peglar over at the Xiotech blog did a really great job of describing “Controller Feature Creep”.  If you haven’t read it, you should.  Also, David Black over at “The Black Liszt” discussed “Mainframes and Storage Blades” he’s got a really cool “back to the future” discussion. 

Today it appears that application and OS/hypervisor vendors have caught up with all the issues we positioned against just years ago.  Exchange 2010 is a great example of this.  VSphere is another one.  Many application vendors now have native deduplication and compression built into their file systems, COW snapshots are a no-brainer, and replication can be done natively by the app which gives some really unique DR capabilities.  Not to mention, some applications support the ability to migrate data from Tier 1, to Tier 2 and Tier 3 based not only on a single “last touched” attribute, but also on file attributes like content (.PDF, .MP3), importance, duration, deletion policy and everything else without caring about the backend storage or what brand it is.  We are seeing major database vendors support controlling all aspects of the volumes on which logs, tablespaces, redo/undo, temporary space, etc. are held.  Just carve up 10TB’s and assign it to the application and it will take care of thin provisioning and all sorts of other ‘cool’ features.   

At the end of the day the “pain in the DAS” that we knew and loved to compete against is being replaced with “Intelligent DAS” and application aware storage capabilities.  All this gives the end user a unique ability to make some pretty interesting choices.  They can continue down the path with the typical storage array controller route, or they can identify opportunities that leverage native abilities in the application and “Intelligent DAS” solutions on the market today to vastly lower their total cost of ownership.   The question the end user needs to ask is, ‘What functionality is already included in the application/operating system I’m running?’ vs. ‘What do I need my storage  system to provide because my application doesn’t have this feature?’  At the end of the day, it’s Win-Win for the consumer, as well as a really cool place to be in the industry.  Like I’ve said in my “Cool things Commvault is doing with REST”, when you couple Intelligent DAS and application aware storage with a RESTful open standards interface, it really starts to open up some cool things. 2010 is going to be an exciting year for Storage.  Commvault has already started this parade so now its all about “who’s next”. 

 @StorageTexan <– Click on me to follow on Twitter. 

 PS – I’ve added an e-mail subscription capability to my site (as well as RSS feeds).  In the upper right corner of this site you will see a button to sign up.  Each time I post a new blog, you will get an e-mail of it.  Also, you will need to confirm the subscription via the e-mail you receive.

Backup-Restores-Power and cooling oh my!!

In today’s datacenter no matter how much de-duplication, storage-tiering, and archiving companies attempt to throw at an issue, there still seems to be an explosion of information that has to be backed up, protected, restored and archived.   I’ve stopped being surprised each time I’ve asked a customer how their backup window is doing.  It’s always horrible and out of control.  Even with advanced data de-duplication it still surprises me at the responses I get.  Not to mention, most of the time customers are running out of power, cooling as well as rackspace in their datacenters.  All of this becomes a sort of “Perfect Storm” that has the potential to sink the datacenter into a mess of inefficiencies. 

So as I’ve discussed in the past, I get to spend a lot of time architecting solutions.  One of the things I spend a lot of time helping design is solutions to help eliminate performance bottlenecks.  The great news is I feel pretty strongly that we have a solution that is best in the industry.  Imagine if you could eliminate storage as a potential bottleneck as well as reduce your power, cooling and overall carbon footprint with one storage solution?   Awesome right!!!  What if I told you that Xiotech has the fastest, best throughput raid-protected spinning media solution in the market today (by Storage Performance Council SPC-2)?  What if I also told you that not only is it the fastest, but it’s also the most greenest (is that a word?) as well?  You would probably tell me I was full of…well you know.  This 3U storage element packs a wallop of performance.  If you haven’t had a chance, you should check out this recent press announcement.  In it we talk about a single Emprise 5000 having the ability to simultaneously power 750 DVD quality video streams, 25,000 MP3’s or 4 Studio-class movie editing projects.  For those of you that are familiar with this types of performance hogging applications you know that in 3U of space, that’s pretty cool!!  We even mention having the equivalent of operating every movie theater screen in the state of Colorado at the same time from one system.  You know on a side note, after 10 years here at Xiotech – how come I don’t have one in my entertainment center yet!!! Brian Reagan – maybe you can make this happen for me :)

You probably noticed that I haven’t even touched on how we can reduce the carbon footprint!!  We have a cool little feature that is native in our Emprise 5000 product called “PowerNap”.  Not only is it native, but it’s also FREE!!  PowerNap utilizes industry-standard Wake on LAN (WOL) technology.  This gives the end user an incredible ability to power up and down the Emprise 5000 solution via scripts or cron jobs.   

Here is something I like to take prospects through.  Let’s say you run a VTL (or backup to disk process) type solution for backup or archiving.  With PowerNap, you can run a simple Perl or PowerScript, as part of a backup process, to spin up the Emprise 5000.  So you kick off your backups at 6pm, it takes just 60 seconds to bring the Emprise 5000 from 24watts of power, up to the full operating  power draw of around 500 watts.  Impressive right!!  So, during the day the unit stays in a low power draw state, only drawing 24 watts of power. NOW THAT’S GREEN!!!!  Let’s say that during the day you need to do a quick restore.  You run your restore process within your backup software solution and part of the restore process is to run the script to spin up the unit.  Once the file has been restored, the backup application can issue another script to “PowerNap” the unit.  By the way, we have a great “BestPractice Guide”.  If you are interested in it, follow me on twitter and send me a message.  I’ll send it over to you.

Did I mention the Emprise 5000 comes with FREE 5 Year Hardware Maintenance and the PowerNap feature is FREE TOO!!!!  In the “me-too” world of Storage array features and functions, it’s things like PowerNap and blazingly fast performance like this that make me happy to go to work each and every day !!

@StorageTexan

What does the Pacer, Yugo and Arbitrated Loop have in common?

What does the Pacer, Yugo and Arbitrated Loop have in common? You are probably running one of them in your datacenter. 

 
 George Crump recently blogged over at InfoWorld and asked, “do we really need Tier 1 storage”?  It struck me as interesting topic and while I disagreed with his reasons on where he put our solution, I tend to agree that the others mentioned are right where they should be.  In his article he specifically mentions some of the reasons both the monolithic array manufactures as well as the “modular guys” have “issues” and he zeroed in on performance and scalability.  Now his article was speaking about the front end controllers, but I think he missed out on pointing to the backend architectures as well.  I thought this would make a great blog posting :)  As you recall in my  “Performance Starved Applications” blog and my “Why running your hotel, like you run your Storage array can put you out of business” blog  I said that if you lined up the various different storage vendors next to each other about the only difference is the logo and the software loaded on the controllers.     

   
 Did you also know that if you looked behind those solutions you would see a large hub architecture – also known as our dear old friend “Mr. Arbitrated Loop”?  This is like running your enterprise wide Ethernet infrastructure on Ethernet hubs.  Can you imagine having to deal with them today?  For all those same reasons we dropped ethernet hubs like a bad habit, you should be doing the same thing with your storage array manufacturer if they are using arbitrated loops  in their backend storage.  Talk about a huge bottleneck to both capacity as well as performance at scale!!  So what’s wrong with Fibre Channel Arbitrated Loop (FCAL) on the backend?  Well for starters it doesn’t scale well at all.  Essentially you can only reference 126 components (for example a disk drive) per loop.  Most storage arrays support dual loops which is why you typically see a lot of 224 drive solutions on the market today, with 112 drives per loop – approaching the limit and creating a very long arbitration time.  Now, for those that offer more, it’s usually because they are doing more loops (typically by putting more HBA’s in their controller heads) on the backend.  The more loops on the backend, the more you have to rely on your controllers to manage this added complexity. When you are done reading my blog post, go and check out Rob Peglar’s blog post around storage controller “Feature Creep” called Jack of All Trades, Master of NONE !.  At the end of the day the limitations of FCAL on the backend is nothing new.   

  

About 4 years ago we at Xiotech became tired of dealing with all of these issues.  We rolled out a full Fabric backend on our Magnitude 3D 3000 (and 4000) solution.  We deployed this in a number of accounts.  Mostly it was used for our GeoRAID/DCI configuration where we split our controllers and bays between physical sites up to 10Km.  Essentially each bay was a loop all to itself directly plugged into a fabric switch.  Fast forward to our Emprise product family and we’ve completely moved away from FCAL on our backend.  We are 100% FULL, Non Blocking, Sweet and as pure as your mamas homemade apple pie Fabric with all of the benefits that it offers!!       

 My opinion (are you scooting towards the front of your chair in anticipation?) is unless you just enjoy running things in hubs I would STRONGLY advise that if you are looking at a new purchase of a Storage Array you should make sure they are not using 15-year old architecture on their backend !!  If you are contemplating architecting a private cloud, you should first go read my blog post on “Building resilient, scalable storage clouds” and applying the points I’ve made, to that endeavor.  Also, if you really are trying to make a decision around what solution to pick I would also suggest you check out Roger Kelley (@storage_wonk) over at http://www.storagewonk.com/.  He talked about comparing storage arrays “Apples to Apples”  and brought up other great differences.  Not to mention, Pete Selin (@pjselin) over at his blog talked about “honesty in the Storage biz” which was an interesting take on “Apples vs Apples” relative to configurations and pricing.  Each of these blog posts will give you a better understanding on how we differentiate ourselves in the market.         

  

Thanks,        

  @StorageTexan       

 

   

  

 

Why running a hotel like you run your storage array could put you out of business.

<this post was updated on April 2, 2010>

Recently I wrote about why “Cost per raw TB” wasn’t a very good metric for comparing storage arrays.  In fact, my good friend Roger Kelley over at StorageWonk.com wrote a nice blog specifically “Comparing Storage Arrays “apples to apples” .  We don’t say this as a means to simply ignore some of the features and functions that some of the other vendors offer.  It’s just our helpful reminder that there is no “free storage lunch”.

So let me take you on a different type of journey around “cost per raw TB” and “cost per useable TB” and apply it to something outside of technology.  Hopefully this will make sense!!

Let’s assume you are in the market for a 100 room hotel.  You entertain all sorts of realtors that tell you why their hotel is better than the others. You’ve decided that you want to spend about $100,000 for 100 room hotel which averages about $1000 per room.   So, at a high level all the hotels offer that same cost per room.  Let’s call this “Cost per raw occupancy”.  It’s the easy way to figure out costs and it looks fair. 

You narrow down your list of hotels to three choices.  We’ll call them hotel C, hotel N and hotel X.   Hotel C and N have the same architecture, same basic building design, essentially they look the same other than names and colors of the buildings.  Hotel X is unique in the fact that it’s brand new and created by a group that has been building hotel rooms for 30+ years with each hotel getting better and better.  They are so confident in their building that it comes with 5 years of free building maintenance.   

So, you ask the vendors to give you their “best practice, not to exceed hotel occupancy rate”.  Hotel C tells you they have some overhead associated with some of their special features so their number is about 60 rooms that could be rented out at any given time.  The reservation system will let you book an unlimited amount of rooms, but once you get over 60 things just stop working well and guests complain.  Hotel N says they can do about 70 rooms before they have issues.  Hotel X says they have tested at 96 room’s occupancy without any issues at all.  

So, while at a high level hotel’s C, N and X were $1000 a room, after further review hotel C is about $1600 a room, hotel N is $1400 a room and hotel X is $1041 a room.  Big difference!!  Let’s assume each of these vendors could “right size” their hotel to meet your 100 room request but the room cost will stay the same.  So, hotel C would now cost you $160,000, hotel N is $140,000 and hotel X is $104,000.  So that my friend is what I like to call “Cost per useable occupancy” !!

Another way to do this is to have hotel C and N right size down to your budget number based on “cost per useable occupancy”.  If the $100,000 is the most important and you understand that you will only get to rent out 60 or 70 rooms from the other hotels, then you could save money with Hotel X by just purchasing 60 rooms in hotel X.  That would bring Hotel X’s costs down to $60,000 or a nice savings of $40,000!!  The net-net is you get 60 rooms across all 3 hotels but 1 offers you a HUGE savings. 

At the end of the day, as the owner of that hotel you want as many rooms rented out as possible.  The last thing you want to see happen is your 100 room hotel only capable of 60% or 70% occupancy. 

So, if you are in the market for a 100 room hotel, or a Storage Array, you might want to spend a little more time trying to figure out what their best practice occupancy rate is !!  It’ll save you money and heartburn in the end.  

I’ll leave you with this – based on the array you have today, what do you think your occupancy rating would be for your 100 room hotel?  Feel free to leave the vendor name out (or not) :)

@StorageTexan

Cool things Commvault is doing with REST

Cool things Commvault is doing with REST 

I’ve always been a HUGE fan of Commvault.  They just rock.  When I was a Systems Engineer back in Austin in the early 2000’s, I don’t think we had an account that I didn’t take Commvault into to try and solve a customer’s backup issues.  AND WE DIDN’T EVEN SELL COMMVAULT !!!   They had such cool technology that was clearly leaps and bounds above everyone else. Not to mention, they had some really cool people that worked for them as well (Shout out to Jeanna, Joelle, RobK and of course Mr Cowgil).  

Fast forward a few years and the release of Simpana as well as the addition of native DeDuplication clearly gave Data Domain and various other deduplication solutions a run for their money.  You would think that would be enough for one company!!   I was pretty excited about their recent press release around adding cloud data storage as a tier option in Simpana.  Dave Raffo over at SearchDataBackup.Com did a really nice job of summarizing the announcement.  It’s a clear sign that Commvault is still very much an engineering driven organization.  Which is just AWESOME!! 

 I think the biggest nugget that I pulled out of the press release is Commvault’s ability to integrate native REST capabilities.  The more and more I hear about REST’s potential, the more I get excited about some of the endless possibilities it can offer.  In this case, it allowed Commvault to easily integrate their backup architecture to include 3rd party cloud solutions like Amazon S3, EMC Atmos and a slew of others.   They didn’t need to build an API for each vendor; they just relied on REST’s open API to do that for them. 

If you haven’t had a chance you should check out Brian Reagan’s blog posting that mentions something we are calling CorteX.  Essentially CorteX is our RESTful based ecosystem on which developers can gain access to our Emprise solutions.  This is the next evolutionary step in our ongoing open architecture capabilities.  As some of you are aware, we’ve been touting our WebService’s Software Development Kit for some time.  It’s allowed us to do things like VMWare Virtual View which ties directly into Virtual Center to give VMWare Admin’s unprecedented abilities, as well as Microsoft developers creating a provisioning application called SANMAN that integrates some of their processes directly to our storage.  RESTful API will take this to a greater level.  Just like Commvault was able to tie directly into public cloud storage providers, CorteX will give unprecedented abilities to do really cool things. 

I’ve probably said more then I should :) So I’ll leave it with “more to come on CorteX as we get ready to release”.  I’ve probably stolen enough of Brian’s thunder to get an e-mail from him!!  It’s always good to hear from a Sr VP right!! 

So, keep an eye on Xiotech over the next couple of months and start paying attention to vendors that support RESTful API’s !!!

Thanks,

@StorageTexan

VMware Virtual Desktop Infrastructure (VDI) and Why Performance matters

Is Storage Performance Predictability when building VMWare Virtual Desktop (VDI) Storage Clouds important?  This can also apply to Citrix and Microsoft Windows Hyper-V Virtual Desktop Systems.

Here is yet another great example of why I just love my job.  Last week  at our Xiotech National Sales Meeting we heard from a net-new educational customer out in the western US.  They recently piloted a VDI project with great success.   One of the biggest hurdles they were running into, and I would bet other storage cloud (or VDI specific) providers are as well, is performance predictability.  This predictability is very important.  Too often we see customer focus on the capacity side of the house and forget that performance can be extremely important (VDI boot storm anyone?).  Rob Peglar wrote a great blog post called “Performance Still Matters” over at the Xiotech.com blog site.  When you are done reading this blog, head over to it and check it out :)

So, VDI cloud architects should make sure that the solution they design today will meet the requirements of the project over the next 12 months, 24 months and beyond.  To make matters worse, they need to consider what happens if the cloud is 20% utilized or if/when it becomes wildly successful and utilization is closer to 90% to 95%.  The last thing you want to do is have to add more spindles ($$$) or turn to expensive SSD ($$$$$$$$$) to solve an issue that should have never happened in the first place.

So, let’s assume you already read my riveting, game changing piece on “Performance Starved Applications” (PSA). VDI is ONE OF THOSE PSA’s!!!  Why is this important?  If you are looking at traditional storage (Clariion, EVA, Compellent  Storage Center, Xiotech Mag3D, NetApp FAS) arrays it’s important to know that once you get to about 75% utilization performance drops like my bank account did last week while I was in Vegas.  Like a freaking hammer!!  That’s just HORRIBLE (utilization and my bank account).  Again you might ask why that’s important?   Well I have three kids and a wife, who went back in to college, so funds are not where they should be at…..oh wait (ADD moment) I’m sure you meant horrible about performance dropping and not my bank account.  So, what does performance predictability really mean?  How important would it be to know that every time you added an intelligent storage element (Xiotech Emprise 5000 - 3U) with certain DataPacs you could support 225 to 250 simultaneous VDI instances (just as an example) including boot storms?  This would give you an incredible ability to zero in on the costs associated with the storage part of your VDI deployment.  This is especially true when moving from a pilot program into a full production roll out.  For instance, if you pilot 250 VDI instances, but you know that you will eventually need support for 1000, you can start off with one Emprise 5000 and grow it to a total of four elements.  Down the road, if you grow further than 1000 you fully understand the storage costs associated with that growth, because it is PREDICTABLE.

What could this mean to your environment?  It means if you are looking at traditional arrays, be prepared to pay for capacity that you will probably never use without a severe hit to performance.  What could that mean for the average end user?  That means their desktop boots slowly, their applications slow down and your helpdesk phone rings off the hook!!  So, performance predictability is crucial when designing scalable VDI solutions and when cost management (financial performance predictability) is every bit as critical.

So if you are looking at VDI or even building a VDI Storage Cloud then performance predictability would be a great foundation on which to build those solutions.  The best storage solution to build your application on is the Xiotech Emprise 5000.

Thanks,

@StorageTexan

Cost Per TB

Greetings and welcome back to my blog !!  I think i’m starting to get the hang of this !!  Hopefully you will agree !!

In my blog post titled Performance Starved Applications we specifically zeroed in on those storage array vendors whose best practice is to keep their systems below a certain utilization rate for performance reasons. If you recall, I mentioned that in our Emprise product line, using ISE technology, we do not have these issues.  Not only can you use more than the industry standard 75%, WE ENCOURAGE YOU TO RUN IT AT 95%% FULL!!!!  Heresy I know!!  We are such rebels over here!!

Let me give you some more food for thought on why this should be important to you.  Why should you pay for the 25% capacity you don’t use?  This is only a big deal if you don’t have an unlimited budget :).  Many times a prospective storage buyer just zero’s in on the upfront costs which in most cases is really misleading.  I like to look at this in a sort of “Cost vs. Value” thing if that makes any sense.   If you’ve ever considered “Cost Per raw TB” as one of the metrics towards the purchase of an array, you want to read further!!

Full disclosure here; I “borrowed” ALL of the following information from Roger Kelley, Principal Storage Architect for the Xiotech Western Region.  He put the “Rock” in Rock Star and I’m not above “borrowing” his ideas!!  His are ALWAYS better than mine anyway!!!  Don’t they say stealing is the perfect form of flattery?  I’m pretty sure that’s how it goes :-)

Cost vs. Value

If I offered to sell you a 100TB SAN for $1000 how many systems would you buy?  Would you consider $10.00 per TB to be a good deal?  I’m mean COME ON it’s freaking $10.00 per TB!!  What if you bought it, got it to your datacenter, installed it, powered it on, opened up your server and found out that all but 1GB of capacity was used for other things?  The upfront cost maybe AWESOME but it’s far from a good value right?  If you can only use 1GB of that 100TB solution you are actually paying $1,000.00 per GB or $1,024,000 per TB!!!  OUCH.

We at Xiotech have referred to this as the industry’s “Dirty little secret”.  This secret is all about highlighting the upfront cost based on “raw” capacity which makes their configs look great.  As you can see from the example above, that doesn’t make this a good value.  Upfront cost per raw TB, in my opinion is not a good thing to focus on.  In my PSA blog we talked about the common industry “best practice” usable being about 75% right? Probably one of the best discussion on this dirty secret happened over at Chuck Hollis Blog back in 2008.  He basically had a “Slap Fight” comparing his EMC (CX4=70%) with NetApp (FAS=60%) and HP (EVA=70%) on their various utilization rates. (STOP don’t leave yet – finish my blog, comment at the bottom how great this was and then go to his and reference my blog while you leave him a comment).  The comments are the place you want to spend your time in.  It’s just a classic slap fight.  You should bookmark it for future reference.  If I can take a moment for some gratuitousness I just want to do a little SHOUT OUT to Chuck Hollis.  I want to be you and Rob Peglar when I grow up!!

Again, I’m not trying to pick on anyone in particular.  This could be any one of the following solutions (or all of them): Clariion, EVA, FAS, Compellent Storage Center, Xiotech Magnitude 3D, etc.  These solutions are typical and represent a vast number of installed arrays.  At the end of that day it’s important to not get wrapped around the axel on “cost per RAW capacity” but to zero in on “cost per useable capacity”.    Unless you just like paying more money then you should!!

Thanks,

@StorageTexan

Performance Starved Applications

In my role here at Xiotech I get to spend a lot of time with prospects as well as customers. It’s probably why I love my job so much. My wife seems to think it has more to do with my A.D.D. (Attention Deficit Disorder) and the ability to change topics and discussions on a daily basis then it does anything else. She’s probably right; at least she likes to point out just how right she typically is!! But that’s probably a separate blog posting :)  So back to my role and the blog topic at hand.

Have you heard of the phrase “Performance Starved Applications” or PSA? Hopefully this term is not new to you. PSAs are simply applications that are not performing correctly due in large part to two things. First, the customer purchased a solution that the vendor failed to size correctly; or if it was sized properly, the customer failed to understand the caveats once it was deployed. What sort of caveats might this be? In typical SBOD (Switched Bunch of Disks) arrays, as the storage system fills up, the performance drops like a hammer. This isn’t just a one vendor issue. Have you noticed that a lot of storage vendors use the same parts to build their arrays? For example, place a Clariion, EVA, FAS, Storage Center, Magnitude 3D next to each other. About the only difference is the logo and the software loaded on the controllers. In most cases they are using the exact same drive bays, shuttles and drives. So, they all suffer from the same issue and they all have a similar caveat when it comes to just how full you can load up their system. The industry has settled on about 75%.

So hopefully we all agree, as the storage system fills up, the performance goes down at just about the same interval. To give you an example, a typical enterprise-quality Seagate drive can produce about 300 “typical” IOPS empty. At 50% full, it drops to about 50% of the performance or 150 IOPS, at 75% its closer to 119 IOPS. So, if we can agree on these numbers the rest should make sense. Let’s play with some numbers. So a customer needs 20TB’s of 15K enterprise drives and let’s say we use those new spiffy 300GB 15k Enterprise Seagate drives. Using my “South Texas” math that’s 20TB/300GB drives = 67 Drives. (before you say it, yes I know a 300 GB drive is NOT 300GB but that’s another blog topic all together – not to mention all the sparing needed for that number of spindles, RAID overhead, etc.) so if we take 67 drives times 300 IOPS we get 20,100 IOPS. SWEET. But let’s be realistic, they have no intention of keeping their storage solution empty. So, should we use 50%? That drops the number down to 10,050 IOPS. Not bad, but let’s be really serious here, NO WAY does a customer only use 50% of their capacity right? Most of my customers use 90% of their capacity but we’ll settle in on the industry standard 75% so that IOP Pool would be around 8000 (just in case you were curious on the 90% number it’s 105 IOPS per drive so that would be 7,000). So I put this into a nice graph because I’m a big fan of pictures !!

<Click to make larger>

So, that’s eye opening.

You essentially paid for 20,100 IOPS but you only get to use 8000. Now think of some of your transactional applications you are running like E-mail and databases and ask yourself “How are they running?” It wouldn’t surprise me if you start to remember a user here or there mentioning that at times their email just sits and waits or database reports that you need take hours to run. Those are what we like to call PSA. And truthfully, most of the time the storage array is the last thing people think of.

 So, how can you solve this issue? The typical answer from vendors is “Buy 50% more capacity”. Problem is, you bought more than you need and never use it :) instead, check out one of our Emprise solutions. They are built on our patented (80+ patents) storage system called the Intelligent Storage Element (ISE). The ISE suffers from no such performance degradation. You can check out a couple of YouTube video’s here – http://www.youtube.com/XiotechCorporation

So, the next time you think your applications are running slow or your end users complain of performance problems, you might want to check and see just how full your array is. If you enjoy using 100% of both the performance and capacity you purchased then give Xiotech a call !!!