Database benchmarks are important tools to assist buyers in deciding which database system is best suited to their needs, but buying a database goes much further than who is the fastest, says Henry Adams, country manager of InterSystems.
A recent report from Klas Enterprises, a research and consulting firm that specialises in monitoring the performance of healthcare IT vendors, compared Oracle`s relational database to InterSystems` object-oriented database, Cach'e. The results created a stir in the ranks of the world`s second largest database company.
For the report: "Application/Database Performance: A Comparison of Applications Based on Cach'e and Oracle", Klas contacted 110 organisations that use four well known applications, two based on InterSystems` database products and two based on Oracle`s database products. The results, without any additional comments or confusing over-analysis, follow below.
Overall, InterSystems rated better in all measures than Oracle did. More specifically, InterSystems rated better than Oracle in every satisfaction measurement, with the biggest Cach'e (InterSystems` object-oriented database) advantage appearing in the area of reliability. When asked about the duration of their last unscheduled downtime, 60% of Cach'e users reported that they had no unscheduled downtime - more than triple the rate (18%) for Oracle users. At the other end of the spectrum, 31% of Oracle users reported that their last unscheduled downtime lasted more than two hours, compared to 13% for Cach'e. The report notes: "The reliability gap between Cach'e and Oracle was the biggest difference measured by this study."
Klas also asked users to rate their satisfaction with the speed and scalability of the database used for their healthcare applications. Again, Cach'e scored significantly higher with 8.12 versus Oracle`s 7.41 (on a scale of 1 to 9).
Despite the fact that Cach'e delivered significantly better performance with a significantly more demanding workload, it also proved to be more economical. To measure this, Klas measured the number of concurrent users per database server and the number of database administrators (DBAs) needed to manage the database.
Cach'e users reported a 2.1 to 1 advantage in terms of hardware required and a 2.5 to 1 advantage in hard-to-find and expensive DBAs. The InterSystems database boasts 465 concurrent users per database server to Oracle`s 219; and 823 concurrent users per DBA to Oracle`s 325.
Klas summarised the results as follows: "The true measure of a solution is its performance in live operation. When we cut through all the hype about technology, state-of-the-art solutions, object-oriented systems, relational databases and client-server, we can see that InterSystems` databases continue to serve the healthcare community as a core component in the best rated systems in the market."
As satisfying as these results are, we must remember that they pertain to the specific needs of the healthcare market where InterSystems has decades of experience. That being the case, the study is also a guide as to the lower cost of ownership an entrepreneurial company can deliver.
For example, global IT services company SchlumbergerSema teamed up with InterSystems to develop CIMS (Consolidated Information Management System), an SMS tracing and statistical measurement application for cellular network providers. "With Cach'e the hardware required for CIMS cost between 5% and 10% of the price of its SMS hardware. On comparable hardware, the Cach'e-based engine is between 10 times and 15 times faster than traditional database systems," Philippe Caille, chief technical officer of the South African subsidiary of SchlumbergerSema, has said.
In addition, companies using Cach'e find that their programmers are much more productive because of the object paradigm. Everything is an object and can be accessed directly, unlike a two-dimensional database which would require normalisation of the query before it could be carried out. In other words, the extent of middleware problems usually associated with database applications decreases substantially.
It is also fallacy to argue that the database is irrelevant anymore. That would be tantamount to saying the foundations of a building are irrelevant as long as the CEO`s suite looks good. But a database can never be measured on its own, which is why Klas measured the performance and costs of applications running on the databases.
A recent Aberdeen report states that new database applications will allow the database market to grow to more than $10 billion by 2003. The company also predicts significant shifts in IT buying and database technology improving the range of choices for buyers across all sectors of the market. "A strong focus on costs, combined with a realisation that, in most cases, it is impossible to combine and standardise databases has led buyers to undertake a more careful assessment of database purchases," says Wayne T Kernochan, Aberdeen managing VP, databases, development environments and software infrastructure."
The decision-making process can be complex, but the rewards of choosing the appropriate database can lead to substantial improvements in the costs of administration - often the most expensive component of a database. In addition, selecting the correct database will also deliver improvements in the performance of applications running on the system and measurable increases in programmer productivity."
Share