Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase 12»»

Anyone Using or Planning to Use Memory-Optimized Tables? Expand / Collapse
Author
Message
Posted Monday, April 28, 2014 10:12 AM
Hall of Fame

Hall of FameHall of FameHall of FameHall of FameHall of FameHall of FameHall of FameHall of FameHall of Fame

Group: General Forum Members
Last Login: 2 days ago @ 3:34 PM
Points: 3,109, Visits: 11,516
I was just wondering if this feature will be used in the real world.

Not being able to have identity columns and foreign key constraints looks like a major drawback, so I wonder if it will be used for anything besides generating impressive TPC tests for Microsoft.



Post #1565643
Posted Monday, April 28, 2014 12:58 PM
SSCrazy

SSCrazySSCrazySSCrazySSCrazySSCrazySSCrazySSCrazySSCrazy

Group: General Forum Members
Last Login: Today @ 1:25 PM
Points: 2,420, Visits: 6,735
I do have plans for implementation but still kind of struggling how to fit this into the architecture. There are certain benefits but the question is more how to fit these into the big picture with a tangible ROI. As it stands, I'm looking at a least few months of parallel operation before I'll make any comments. As you mentioned, there are drawbacks and at least from my standpoints, those are significant.
Post #1565679
Posted Monday, April 28, 2014 2:50 PM


SSC-Dedicated

SSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-Dedicated

Group: General Forum Members
Last Login: Today @ 5:25 PM
Points: 35,615, Visits: 32,210
Michael Valentine Jones (4/28/2014)
I was just wondering if this feature will be used in the real world.

Not being able to have identity columns and foreign key constraints looks like a major drawback, so I wonder if it will be used for anything besides generating impressive TPC tests for Microsoft.


For the very reasons you mention and more, no... I have no plans to use Hekaton for at least the foreseeable future.


--Jeff Moden
"RBAR is pronounced "ree-bar" and is a "Modenism" for "Row-By-Agonizing-Row".

First step towards the paradigm shift of writing Set Based code:
Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column."

(play on words) "Just because you CAN do something in T-SQL, doesn't mean you SHOULDN'T." --22 Aug 2013

Helpful Links:
How to post code problems
How to post performance problems
Post #1565711
Posted Tuesday, April 29, 2014 6:11 AM


SSChampion

SSChampionSSChampionSSChampionSSChampionSSChampionSSChampionSSChampionSSChampionSSChampionSSChampion

Group: General Forum Members
Last Login: Today @ 8:42 AM
Points: 14,037, Visits: 28,407
The thing is, I think the technology is amazing, but, it's really all about the ability to collect data and feed data out at high speed in an OLTP environment. Way too many people are going to think that it solves all sorts of problems that it doesn't. It's not a panacea. In fact, it's extremely focused in the problems it can solve. But, I think it solves those problems well. But, again, you have to have big enough machines to really make it work. We're still getting all sorts of questions from people running on an 8gb server. They're not going to implement in-memory anything any time soon. It's a niche solution, albeit a pretty cool one. But no, no immediate plans because we're no where near the target audience and don't have the kind of hardware necessary to make the whole thing work.

----------------------------------------------------
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood..." Theodore Roosevelt
The Scary DBA
Author of: SQL Server Query Performance Tuning
SQL Server 2012 Query Performance Tuning
SQL Server 2008 Query Performance Tuning Distilled
and
SQL Server Execution Plans

Product Evangelist for Red Gate Software
Post #1565873
Posted Tuesday, April 29, 2014 7:26 AM
Forum Newbie

Forum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum Newbie

Group: General Forum Members
Last Login: Thursday, November 6, 2014 10:46 AM
Points: 1, Visits: 95
No plans here so far. We looked into it, the problem would be there is not a way to add a where clause for the memory table. (hint: Microsoft). Our problem is we would be dealing with mass amounts of data, but we would only be interested in say the most current 3 months transactions for a particular customer as opposed to dealing with the entire customers history.
Post #1565910
Posted Tuesday, April 29, 2014 7:40 AM


SSC-Forever

SSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-ForeverSSC-Forever

Group: General Forum Members
Last Login: Today @ 8:31 AM
Points: 40,456, Visits: 36,912
No plans, but I have advised two clients to consider replacing SQL 2012 with SQL 2014 in their upgrade plans. Partially because, knowing these guys, they won't actually upgrade until around 2016, but also so that I can consider Hekaton for very specific, very carefully targeted portions of their systems.

Not that they're in any way at the point that there are no other performance gains to be had. Hell, most databases I see it's a combo of 'Did a cat throw up on that stored proc, or did you write it that way?', 'Indexes? What are they?' and '2003 called and would like its hardware back'



Gail Shaw
Microsoft Certified Master: SQL Server 2008, MVP
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability

We walk in the dark places no others will enter
We stand on the bridge and no one may pass

Post #1565922
Posted Tuesday, April 29, 2014 7:44 AM


SSC-Dedicated

SSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-DedicatedSSC-Dedicated

Group: General Forum Members
Last Login: Today @ 5:25 PM
Points: 35,615, Visits: 32,210
Grant Fritchey (4/29/2014)
The thing is, I think the technology is amazing, but, it's really all about the ability to collect data and feed data out at high speed in an OLTP environment. Way too many people are going to think that it solves all sorts of problems that it doesn't. It's not a panacea. In fact, it's extremely focused in the problems it can solve. But, I think it solves those problems well. But, again, you have to have big enough machines to really make it work. We're still getting all sorts of questions from people running on an 8gb server. They're not going to implement in-memory anything any time soon. It's a niche solution, albeit a pretty cool one. But no, no immediate plans because we're no where near the target audience and don't have the kind of hardware necessary to make the whole thing work.


I absolutely agree. It is both a niche and nice solution for high speed OLTP and a lot of people are thinking of it more as a general purpose RAM disk, which it isn't. It's very targeted and does well for what it was designed for. I do wish it had IDENTITY column capabilites, though.


--Jeff Moden
"RBAR is pronounced "ree-bar" and is a "Modenism" for "Row-By-Agonizing-Row".

First step towards the paradigm shift of writing Set Based code:
Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column."

(play on words) "Just because you CAN do something in T-SQL, doesn't mean you SHOULDN'T." --22 Aug 2013

Helpful Links:
How to post code problems
How to post performance problems
Post #1565929
Posted Wednesday, April 30, 2014 9:12 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Sunday, May 4, 2014 10:21 PM
Points: 199, Visits: 307
From what I understand, if you convert any table to use the clustered columnstore format, you're implicitly using this. It stores the first 10k rows (I believe) in the in-memory OLTP table then uses a tuple mover to move it to the columnstore. Hence its "updatable".


Post #1566453
Posted Wednesday, April 30, 2014 12:45 PM


Old Hand

Old HandOld HandOld HandOld HandOld HandOld HandOld HandOld Hand

Group: General Forum Members
Last Login: 2 days ago @ 7:02 AM
Points: 372, Visits: 981
No. I did some testing with it and found the performance increase to be very nice. However, in order to get the tables involved to comply with requirements they basically had to be modified in many ways. That would mean the application(s) would have to be modified as well. None of that is going to happen anytime soon. There is also the question of memory for the server itself. In order to really load up we would need to move to the Enterpri$e Edition and the cost are just not something we are going to deal with at this point. Things like a SAN are way ahead of that.

Cheers
Post #1566554
Posted Thursday, May 1, 2014 5:54 AM
SSC Eights!

SSC Eights!SSC Eights!SSC Eights!SSC Eights!SSC Eights!SSC Eights!SSC Eights!SSC Eights!

Group: General Forum Members
Last Login: Monday, November 24, 2014 12:25 PM
Points: 969, Visits: 974
Anyone think that a good use case for a Hekaton table would be for the ASPSession database tables? Obviously there are alternatives to caching session data, but for those using SQL, it might be a good solution (if the FK's and identity cols don't exist.. not on my work computer at the moment to check for those in that DB).
Post #1566679
« Prev Topic | Next Topic »

Add to briefcase 12»»

Permissions Expand / Collapse