SPSS Stat Server - advantages vs. SAS, other

[Steve Few]

On platform/user purchases of SPSS / PASW modules:

What are you using? I need to upgrade.I am using SPSS v17 at the moment. I
am having some limitations in-memory which causes problems.This leads me
to PASW + Stat Server. I think I am looking at $$. Perhaps 3 users.

Does anyone have comments on how Stat Server may benefit me given I do
have to share code and data in the future in my group? Compared
conceptually to SAS enterprise guide on 3 desktops while using a MS
Server? How about RStat, open source? I've used R, and doing so now in
special cases.

METHODS: - Advanced Modeling Time Series, Categorical, Classification,
Linear models, Mining, etc.

Two issues to consider:
1. Server, multiple users, it's specific data handling on th server.
Common datasets, good idea. Downside might be cost.
2. Advanced models that might run better if handled by a client desktop,
hitting the Stat Server. It seems the memory handling should be superior.
What is your experience? Suggestions?  Facts?


I have contacted technical sales, but no info yet. Just thought I would
get experienced user's view(s).

Thanks in advance, I will share results of my decision.
Steve

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Rangel, Aaron]

Hi Steve,

I am a Product Manager on the Statistics team responsible for Statistics server. Here are a few reasons why customers buy Statistics server, I understand all of these may not apply to you,

1.      Working with large files. Basically doing data transformations on large files, sort, merge... If you dealing with large files (above 2GB) these data transformations can be time consuming and lock up the machine. Data corruption is also a frequent complaint.

2. Working with large files across offices. Think of an organization that is centralized in Texas (All the large datasets are in the Texas datacenter along with the SQL server) with offices in San Diego and New Jersey. With Statistics server, users in the distributed offices can conduct analysis without the need to transfer the large datasets across offices over the WAN (wide area network). WAN's are much slower when compared to LANs and the efficiencies realized in terms of time saved because these files are not transferred across offices especially when the files are large are very significant.

3. For the reasons listed in point 3, Statistics server is ideal for an organization with distributed users

4. Faster performance, this is because we don't limit the number of threads on Statistics Server. This allows Statistics server to make maximum use of available CPUs for multi-threaded procedures. I have listed some of the multi-threaded procedures
Procedure family Procedure Name
Correlations Bivariate, Partial
Regression Linear,Ordinal,Multinomial,Logistic
Data Reduction Factor Analysis
Survival Analysis Cox Regression
Multiple Imputation Impute missing values
Sort

5. Statistics Server facilitates better security and standardization. This is because datasets and syntax are located in a central location vs. on users desktops and laptops

6. Statistics Server is also a foundational technology that you can leverage in several ways. When combined with Collaboration and deployment services, an organization can deploy simple web interfaces for statistics capabilities in a portal environment to business users in an organization

Hope this helps, feel free to contact me at [hidden email] to discuss further. I can also put you in touch with customers who are using Statistics Server successfully so you can better determine whether Statistics Server is right for you.

Aaron





-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Steve
Sent: Monday, August 10, 2009 1:31 PM
To: [hidden email]
Subject: SPSS Stat Server - advantages vs. SAS, other

On platform/user purchases of SPSS / PASW modules:

What are you using? I need to upgrade.I am using SPSS v17 at the moment. I
am having some limitations in-memory which causes problems.This leads me
to PASW + Stat Server. I think I am looking at $$. Perhaps 3 users.

Does anyone have comments on how Stat Server may benefit me given I do
have to share code and data in the future in my group? Compared
conceptually to SAS enterprise guide on 3 desktops while using a MS
Server? How about RStat, open source? I've used R, and doing so now in
special cases.

METHODS: - Advanced Modeling Time Series, Categorical, Classification,
Linear models, Mining, etc.

Two issues to consider:
1. Server, multiple users, it's specific data handling on th server.
Common datasets, good idea. Downside might be cost.
2. Advanced models that might run better if handled by a client desktop,
hitting the Stat Server. It seems the memory handling should be superior.
What is your experience? Suggestions?  Facts?


I have contacted technical sales, but no info yet. Just thought I would
get experienced user's view(s).

Thanks in advance, I will share results of my decision.
Steve

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD