BookmarkSubscribeRSS Feed
Ram4sas
Fluorite | Level 6

Hi,

 

I would like to know how the forecast studio works in the back groud like

when we launch Forecast studio client(PC client) and run project which authenticate with metaserver for log in and with help of Metadata connection would connect to SASApp server components and data.

 

what exactly web server role when run the forecast studio project on forecast studio client(PC)?

what exactly web server role when run the forecast studio project on forecast studio client(web launch app)?

 

if run the forecast project in forecast srudio client, what is the normal response time when select different row/variable data under table view and for different model under modeling view? like how long it takes to display result for each selection?

 

if forecast studio project run and taking 3-4 seconds for display results for each selection does it happening cause of web server JVM settings?

 

what are recommend JVM valus for -Xms and -Xmx /Permsize and MaxPermsize? does this really has impact on forecast server response activity? 

 

 

Thanks,

Ram

 

1 REPLY 1
rajib_nath
SAS Employee

Hi Ram,

 

See my response below:

 

what exactly web server role when run the forecast studio project on forecast studio client(PC)?

what exactly web server role when run the forecast studio project on forecast studio client(web launch app)?

 

The middle tier is an interface between user input (through some client) and the actual SAS execution. It primarily interpretes the instructions from the client and creates instruction which SAS understands. There is not much difference between the way FS client and FS Web client (through webstart) operates. The client has to be installed in a desktop machine (a thick client) where as FS JWS client is sort of a thin client. The client pieces are downloaded in the runtime and someone can use forecasting even without installation.

 

if run the forecast project in forecast srudio client, what is the normal response time when select different row/variable data under table view and for different model under modeling view? like how long it takes to display result for each selection?

 

There is no fixed answer to this - depends on the size of the data being request. If the graph / trable has a large number of data points, it will take longer to render the graph. Depending on the view you are into, the technology also differs, which would also vary the rendering time. The Forecasting view uses silk graph and modeling / timeseries view uses statgraph. Mpodeling view has additional overhead of loading the models increasing the data transfer between server to client. Network latency and all such factors would potentially impact rendering time of plot / results in various views. Sorry, i could not give a concrete answer to your question.

 

if forecast studio project run and taking 3-4 seconds for display results for each selection does it happening cause of web server JVM settings?

 

Well, if you are experiencing delay on loading plots, there can be multiple reasons - client memory settings, network latency, server slowness etc. I would guess you can improve the experience by increasing memory settings in your client first:

 

You can attempt the same scenarios by adjusting Xms, XmX settings for client. These modifications can be made under 'forecaststdo.ini' file, which can be located inside '<SASHome>\SASForecastStudio\<version>\'

 

If you are interested in tuning overall performance, you can experiment by changing other parameters like Xms, XmX, permSize etc. in the middle tier. Server size performance depends on MEMSIZE, SORTSIZE settings. 

 

what are recommend JVM valus for -Xms and -Xmx /Permsize and MaxPermsize? does this really has impact on forecast server response activity? 

 

I would say the recommended settings are what is specified by default. The usage pattern varies from user to user, so additional tuning may be required.

 

Please feel free to reply for any doubts.

 

Regards,

 

 

 

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

Multiple Linear Regression in SAS

Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 1 reply
  • 875 views
  • 0 likes
  • 2 in conversation