CASE STUDY - Data Warehouse Appliances in High Tech Manufacturing
By Mark Theissen, VP Professional Services, DATAllegro
In order to remain competitive, high-techmanufacturing companies mustmanage their costs, especially theirIT costs. The following scenario wasrecently faced by a leading high-techmanufacturer in the computing industry;we’ll call the company HTM, Inc.
HTM had great success in their implementationand use of data warehousingto improve production and quality anddrive competitive advantage. The successof their data warehousing effortsresulted in increasing numbers of usersand increasing demands for data. Being aglobal company, HTM’s data warehousewas accessed worldwide. Batch windowsfor daily updates were becoming smallerwhile the workload was increasing. Thedata itself continued to grow as productiongrew at double-digit rates. Inaddition, users were demanding thatmore and more data be retained for historicalanalysis. The data warehouse wasreaching its capacity.
The first option HTM explored was expanding their existing infrastructure to meet increasing demands. But the existing infrastructure was expensive, and it became apparent that the price tag associated with increasing the size of the data warehouse environment would be costprohibitive. HTM needed an alternative that would add more capacity, increase scalability, and reduce costs instead of increase them.
As a leader in high technology, HTMbegan to look at emerging technologiesto meet its data warehouse infrastructureneeds. Data warehouse appliances quicklybecame a point of focus, because theyoffer a number of benefits:
- A purchase price close to what HTM had been paying annually for the maintenance of its existing data warehouse infrastructure.
- Increased query performance over the existing environment.
- Load times that were hundreds of times faster than current methods.
- Room for growth and the ability to expand the appliance without replacing it.
- A lower total cost of ownership.
All of these benefits translated intoan attractive solution that needed to beproved out before it could be purchased.A proof-of-concept project was runthat exercised the appliance in terms ofthroughput, concurrency, and scalabilityagainst a mixed workload of queries. Theresults were outstanding, with load timesover 200 times faster than the existingenvironment. Queries ran from 10 toover 100 times faster than the existingdata warehouse.
Best yet was the fact that implementingand maintaining the appliance wasa straightforward process. The infrastructure(i.e., database, hardware, OS,software, and storage) was pretuned andconfigured for data warehousing. Theexisting data warehouse data model couldbe used with no modifications. DBA andsystem administration time was reducedas the number of indexes was minimized,space management was automated, aggregationswere not required, and querytuning was greatly reduced.
Today, many companies have the same challenges that HTM faced. Usage of the data warehouse is increasing, the data warehouse is becoming mission critical, and the variety, velocity, and volume of data is increasing. Data warehouse appliances can drive IT costs down while delivering new levels of performance and scalability at a price that cannot be matched by traditional data warehouse infrastructures. Indeed, data warehouse appliances can enable analytics that were previously impossible or unaffordable.
It should be noted that the query results from the existing system were run while the system was running concurrent processes, and the queries run in the data warehouse appliance environment were run individually. That said, the queries on the data warehouse appliance were run without any tuning or indexing, while the queries on the existing system had been tuned and optimized by a team of people over several months.
This article originally appeared in the issue of .