Re: VLDM Slow performance on Client PC

On 7 May, 15:48, "Jeje" <willg...@xxxxxxxxxxx> wrote:
maybe....

try to change the connection string and play with these parameters:
LArge Level Treshold=<number>

if the number of members is higher then the number then the query is
executed on the server side. This impact the way the members are retrieve
from the server to the client.

you can force the execution of the query on the server side rather the
client side with these parameters:
Cache Policy=7
Execution Location (I don't remember the value for this property, search in
google to found it)

I hope this will help you.
good luck.

Jerome.

<kis...@xxxxxxxxxxxxxxxxx> wrote in message

news:1178401976.724986.276080@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx



Hi,

I am using AS2000 and have a Very Large dimension called Product.
This contains almost 1 million products. As expected when processing
it creates another system process to hold the dimension in memory.
This is fine and I realise that this is a limitation on the 32 bit
version.

The problem is on the client PC, when users begin to query the cube.
Users have Intelligent Apps installed as an add-in for Excel. When
the product dimension is queried, we have noticed that it IA/Excel
begins to download the entire dimension into memory as we can see the
Excel.exe memory size increase to 400MB which is the size of the large
dimension.

The knockon affect is that queries take a long time to run as I
suspect it is trying to do things from local PTS cache on client PC
before going to the server if it can't resolve.

Since the client PC's are not as fast and have limited memory, any
queries we run using product dimension never completes.

Is there anyway we can force all queries to happen on the server and
only the requested results are sent back to the client or is this a
limitation in Excel PTS where it tries to resolve locally?

Any advice would be gratefully apreciated.

Thanks
Kishor- Hide quoted text -

- Show quoted text -

Hi ,

Thanks for the pointers. I have tried the above and lowered the Large
Level Threshold to 100. Execution Location =3 for server but still
the VLDM dimension is downloading to client.
The same view and query run faster on a zero footprint web client. I
read somewhere that connecting to cubes via HTTP provide a better
performance because it goes through IIS we caching. Is this true?

Anyway we are working with Microsoft on this, so will post the outcome

Thanks

.

Relevant Pages

  • Re: High CPU in client (Excel, OWC, Proclarity, etc.) accessing Analysis Services
    ... We have encountered similar problems not limited to Excel, but every client. ... occur when there are many dimensions nested on each other on a single axis ... > Performance Guide to optimize the query, the cube, the server, etc. ... > While the query was grinding, the server was doing absolutely nothing. ...
    (microsoft.public.sqlserver.olap)
  • RE: LDAP query for xerox 3545 printer/scanner blocked by SBS 2003?
    ... Client and printer pointing to SBS for DNS ... LDP query from another workstation was successful ... Step2:Please ensure the DNS on workstation is pointed to SBS server. ... Monitor will capture some packets which pass through this interface. ...
    (microsoft.public.windows.server.sbs)
  • Re: Performance Benchmarks?
    ... adding memory on the server can help you; more data can be cached on the ... Here is my test query. ... )} DIMENSION PROPERTIES MEMBER_CAPTION, MEMBER_UNIQUE_NAME ON ROWS FROM ... I have created aggregations (Partitions tab in cube design). ...
    (microsoft.public.sqlserver.olap)
  • Re: public distributed filesystem
    ... You have a simple server which keeps track of client ... >>whether a peer has honestly kept your data block ... respond to any checksum query is to essentially store ...
    (comp.os.linux.misc)
  • Re: VLDM Slow performance on Client PC
    ... HTTP access is slower than direct access. ... from the server to the client. ... > it creates another system process to hold the dimension in memory. ...
    (microsoft.public.sqlserver.olap)
Quantcast