Desktop productivity for business analysts and programmers

Enterprise, Metadata, SAS Grid

Occasional Contributor
Posts: 11

Enterprise, Metadata, SAS Grid

Can anyone explain in plain English what SAS means by Enterprise Solutions (I always thought Enterprise was a spaceship) and what the heck metadata is and what a metadata server is (do I still use SAS Data Sets?)?

Also, can anyone explain in simple terms and steps how I would set up up a simple SAS grid? I've set up a SUN grid that works well on non-interactive SAS jobs, but I understand the SAS grid will work with the SAS interactive screen. Is that right?

I have 6 SUN workstations and a SUN server. Would the SUN server be the "metadata server"? Where does the actual data go? Do I still use SAS Data sets? Where would they go? What do I need to put on the SUN server? What do I need to put on the SAS workstations that both submit jobs and run them?

I want to do load-balancing and use all 6 workstations.

Trusted Advisor
Posts: 2,127

Re: Enterprise, Metadata, SAS Grid

SAS has pretty good descriptions of these on their support web site.

Cynically, "Enterprise Solutions" seems to mean "Extra Cost Solutions," as they appear to be separately licensed products.
Occasional Contributor
Posts: 11

Re: Enterprise, Metadata, SAS Grid

Maybe I'm just not smart enough to understand all the desciptions SAS supplies:

a description or definition of data or information.

metadata LIBNAME engine
the SAS engine that processes and augments data that is identified by metadata. The metadata engine retrieves information about a target SAS data library from metadata objects in a specified metadata repository.

metadata object
a set of attributes that describe a table, a server, a user, or another resource on a network. The specific attributes that a metadata object includes vary depending on which metadata model is being used.

metadata repository
a collection of related metadata objects, such as the metadata for a set of tables and columns that are maintained by an application. A SAS Metadata Repository is an example.

metadata server
a server that provides metadata management services to one or more client applications. A SAS Metadata Server is an example.

Exactly what does it mean by "a description or definition of data or information"? What kind of desciption or definition is kept and supplied? Is this information about SAS Data Sets? Are SAS Data Sets still used? What does it mean by "a set of tables and colums that are maintained by an application"? None of my programs have tables or colums that I know of. We use straight-forward SAS Data sets. I understand them. Meta data - not so much.

Does anyone understand them enough to explain them in "Plain English"?

Posts: 53

Re: Enterprise, Metadata, SAS Grid

Roger, it takes a while to wrap your brain around this stuff, but imagine that your boss asked you to create a summary file of orders for a new widget. The specifications that he/she provides are the metadata. A good number of the new fangled software that SAS provides deals with metadata and NOT the data itself.
A simple example of metadata is the specs your boss might provide for a summary file of orders for widgets. At the time the boss hands the specs to you, the new file does not exist, but the DESCRIPTION does and SAS tools can manipulate (and provide useful output!) using this data. Hope this helps.
SAS Employee
Posts: 149

Re: Enterprise, Metadata, SAS Grid

Hi Roger,

This is not an official SAS explanation, but here are some basic ideas I have on the subject. The term metadata translates loosely as "data about your data" -- the reason its tricky to define is because its so fundamental to other definitions. It's sort of like defining a "point" in Euclidean geometry. We all know what a point is... but providing a definition that isn't wishy-washy is tough. Maybe the best way to understand metadata is to look at some examples:

Things that are metadata: name, type, and length of each column in a SAS data set; the location of different servers defined for SAS to use; users and groups that have permission to access SAS; names and locations of tables in a SAS folder; libraries defined for SAS to use; system options and settings

Things that are NOT metadata: values in a table... this would be data, not metadata.

Hopefully this helps a little. Metadata is sort of everything you interact with in a SAS platform environment except data values.

A metadata server is what holds all of this cool information in one or more metadata repositories. It keeps everything talking to everything else. When you submit a SAS program for example, the metadata server would figure out if you have the appropriate permissions to access the tables and columns in question, figure out what server the SAS code will execute on, and so forth. The server that does the heavy lifting, i.e. that runs the SAS code, is referred to as the Workspace server.

In a SAS platform environment where you have a metadata server etc., of course there are still SAS data sets. Rather than existing out in some random corner of a hard drive until you submit a LIBNAME statement, SAS data sets in a SAS platform environment are typically registered in the metadata. That means anyone with the appropriate permissions can find them easily, whether the machine they're stored on is 10 feet or 10,000 miles away.

Enterprise Solutions (I believe... not an expert on this) are basically off-the-shelf software products designed for specific business needs. For example, SAS Anti-Money Laundering is an Enterprise Solution we sell. The word "enterprise" means that people across your organization with different backgrounds and business needs can access it. A piece of SAS code which you write would be a "solution" to a problem, but not one that's easily translated into a large-scale application for use across the organization.

For more information, someone in SAS sales, SAS tech support, or SAS education would be a great resource for better understanding how these terms apply to what problems you're trying to solve with you SAS software.

Ask a Question
Discussion stats
  • 4 replies
  • 4 in conversation