By James e. PowellJanuary 8, 2013
What you need to know about the most important topics of BI and DW-master data management for large amounts of data, selection of data virtualization platform? To find out which topics BI and DW professionals should look-and pick up pointers we haven't heard before--we talked about William McKnight, McKnight Consulting Group and an experienced, credentialed and information management success strategist and practitioner.
Mr. McKnight is a keynote speaker at the upcoming TDWI World Conference in Las Vegas, February 17-22. 21 February he will discuss "Capitalizing on the chaos: possibility to manage navigation information to build organizational value.
BI this week: talk a lot of platform selection, but they are not all that so quickly these days that doesn't matter too much which one you choose?
William McKnight: It is very important. Of course, platforms are faster than in the past, but is still behind the requests, especially if the workload is set incorrectly. The main consideration is the category of platform that adapts to the workload. Considering all the options is a must, because we are dealing with the important--arguably the most important these days-the company's asset: information.
Some platforms are genre-bending and, of course, the sellers have blankets until after the sale and then all the bases, "well, you must split the workload and to enter in our offer in this category." Go there first. Give the best workloads. Do not assume the necessity of an implementation of errors, either. The right platform can accommodate the inevitable suboptimal development and tuning, but a poor platform category selection it gives little room for error. We could certainly quibble over the selection of latest technology, but dividing the workload for their characteristics and assigning these workloads to their best platform is a must. Let me repeat that: there is a better platform for each workload. With experience, you can get there quickly.
We should not be cramming master data management (MDM) functionality and large amounts of data into the data warehouse, for example. Nor should we be treating all data access by providing uniform out one as the data warehouse with a single tool.
You are cited around TDWI talking about how frequently the various functions are being pulled out of the data warehouse. With several stores in the analytic mix now, is one called a data warehouse must still exist?
Yes, it does. Take to mean an operational data store that has a primary function of feeding other systems and a secondary function of storing historical data. With regard to serve countless data access mode, not so much. This shop will sit alongside architecture analytic shops and, of course, earlier in the cycle of some analytic shops. There are architectural independent data marts and data warehouses is not necessarily the Sun around which everything orbits. Some of these functions is pulled out of the data warehouse are going well operational. The stream processing and master data management are obvious examples.
0 comments:
Post a Comment