Which hardware and operating system platform(s) would you recommend for placing the actual data warehouse data on (one terabyte initially and then growing)?


Sid Adelman’s Answer: You shouldn't limit your considerations just to the hardware and operating system. You should also consider the database as well. When you do your reference checks, ask for installations which have implemented tables that are at least as large as you intend your tables to be. Ask them what performance problems they have had and how they and the vendors dealt with those problems. Michael Jenning’s Answer: I have experiences with successful 1+ terabyte data warehousing projects on HP/HP-UX, NCR/NCR UNIX and Sun/Solaris. Most hardware and OS combinations today would probably meet this criteria (my home PC almost does!). Some other considerations I would include are:

  • What DBMSs support the selected hardware and OS? Can the DBMS selected accommodate growth into multiterabytes? Some DBMSs have upper growth boundaries once you get beyond 20 terabytes.
  • Is the hardware, OS and DBMS combination selected supported by your preferred data warehouse tool vendors (data modeling, ETL, meta data repository, query and reporting, cleansing, data profiling)? Some combinations may not be supported by your tool vendors, or only supported at a second- or third-tier level, leaving you waiting for new features, or worse, dropped from future support.
  • What hardware, OS and DBMS experiences do you currently have within your firm? There are obviously training, maintenance, support and possibly consulting costs that would be incurred with new items being introduced into your firm’s environment plus a longer implementation period for transition time.
  • Can the hardware, OS, DBMS and network environment be used in conjunction with network attached storage (NAS) or a storage area network (SAN)? Either of these methods are viable considerations for warehouse project’s that need to control cost and/or database growth.

Chuck Kelley’s Answer: This depends on your company’s standards. I would put it on a UNIX box, because I don’t think that Win2000 is quite ready for the 1TB databases and the user community required to justify it (Please no hate e-mail – this is my opinion! I will not be devastated if you think otherwise.).

Clay Rehm’s Answer: When considering data warehouse hardware and software, I always look at and assess what is already at the client site first. What hardware and software has your company standardized on? Will this hardware and/or software work for you? Why not? After you have identified the standards, you can ask the question that just because it is a standard at your company, is it still a viable tool? My advice is that you are always better off using what exists for your initial releases of the data warehouse. This gives you a chance to determine how big the DW will possibly become, how many users will be accessing it and how they may be accessing it. You may make the mistake of purchasing hardware and software early on only to discover it was not enough, it was overkill or it was completely the wrong choice.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access