Posted on 09/28/2010 5:04:24 PM PDT by WebFocus
InformationWeek Analytics, the leading service for peer-based IT research and analysis, today announced the release of its "Research: 2010 State of Database Technology" report. More than 750 business technology professionals weigh in on their database strategies. Report author Richard Winter is founder and president of WinterCorp, an independent consulting firm that specializes in the performance and scalability of data management systems.
Research Summary:
Our first InformationWeek Analytics State of Database Technology Survey reveals serious fault lines beneath the critically important enterprise database and data warehousing markets. The 755 business technology professionals taking part in our poll express discontent with rising license and upgrade fees as database sizes and workloads spiral ever larger. On the vendor side, a spate of acquisitions and the rise of appliances are creating uncertainty as well.
Findings:
* Most respondents, 88%, hail from enterprises where the primary operational database platform is from Microsoft (35%), Oracle (35%) or IBM (18%). While the majority are generally satisfied with features and performance, more than half, 52%, take issue with license fees; 13% of respondents characterize their costs as "highway robbery."
* A remarkably high percentage of respondents, 27%, are using as a secondary operational database the open source MySQL, which is now owned by Oracle and, more importantly, carries no license fee. In addition, 39% are interested in NoSQL, a term encompassing a group of large, clustered but nonrelational data management systems, often inexpensive or open source. Together these trends suggest we'll see movement toward alternatives to the commercially available relational database platforms that have been the near-universal standard for the past 25 years.
* The data warehousing market is also in flux. The good news is that 41% of respondents have a single enterprise data warehouse (EDW) or are working toward that goalthe largest percentage pursuing any single strategy in our survey. However, just over 60% are satisfied with the performance and features of their EDW platforms, while a quarter are unhappy with license fees.
* MySQL is frequently cited as a secondary data warehouse or data mart, and in the analytic databases category, which 48% see as distinct from data warehouses and data marts, a remarkable 22% of respondents are using, experimenting with or investigating the open source platform Hadoop; slightly fewer are looking at related tools, such as BigTable or MapReduce.
"One subject we really dug into for this report is database security, an area often underemphasized," says Lorna Garey, content director of InformationWeek Analytics. "We're worried that respondents are talking a good game here while not always following through. I say this because while 70% say their organizations perform database security assessments to identify weaknesses, only about half of those people were able to name the security assessment products in use. Maybe they're assuming that the CISO is handling this, but the phrase 'trust but verify' comes to mind for me."
InformationWeek Analytics is a subscription-based service, offering peer-based technology research. Its site currently houses more than 900 reports and briefs, and includes a dedicated area where technology professionals can access complete issues of InformationWeek Magazine.
Oracle is only run on the Solaris servers. MYSQL on both.
So, how do you compare MYSQL running on Windows compared to SQL Server? ( What versions of MYSQL and SQL Server are you guys using anyway?)
I believe the article is showing an increased market share for SQL Server because it has been designed and dedicated to the Windows platform.
Like it or not, Windows has over 90% of the OS market worldwide and Microsoft’s dominance in this area will simply ensure that any product they develop, be it Database Servers or Web servers or browsers will be OPTIMIZED for the Windows platform ( whatever version they’re coming up with ).
SQL2005 still annoyed me, especially the full-text indexing as a separate process, very kludgy. SQL2008 is pretty sweet though, I love working with it.
We are running MYSQL 5.5 using myphpadmin to build database on UNIX and MySQL admin to build databases on Windows.
We use these two databases to keep it simple and consistent. The web engineers know there databases well. We have over 100 web sites to keep running.
I generally use one these days, MSSQL. I’ve found over the years that what’s far more important for performance than the database engine or even the hardware you use is the design of the database and how it’s used by the application. I’ve achieved literally 100x+ performance improvements combined with 90% drops in server load by changing even small parts in the structure and access of databases created by others. They thought SQL Server or the hardware was just slow, but it was the design that was at fault. Why throw $50,000 more at hardware and SQL licenses when a few days of intelligent analysis and redesign will fix it?
Once upon a time I acheived a factor of 30 speedup in a kernel for a computational routine at work by re-arranging and re-writing it to eliminate algebraic redundancies.
Cheers!
Once upon a time I achieved a factor of 30 speedup in a kernel for a computational routine at work by re-arranging and re-writing it to eliminate algebraic redundancies.
Cheers!
It would spur otherwise unnecessary purchases of Windows licenses.
But these days, I find no reason to go beyond SQLite, in those cases where I even need a RDBMS. There is no reason to pay for the licenses you need to run MSSQL, much less Oracle. Google doesn't do it, so why should I?
There is if you have data requirements too big for SQLite. It is "lite" after all. It can't really handle highly transactional or large datasets.
Google doesn't do it, so why should I?
Oracle uses a custom, distributed, non-relational database system called BigTable that has a multi-tiered lookup architecture and is designed to work on top of Google's custom file system. It's designed for fast lookup within petabytes of data across thousands of servers. It's also not for sale.
Why did I write Oracle? BigTable is of course Google.
I know BigTable isn't relational. As for MySQL, it has definitely come a long way. It couldn't even do transactions, stored procedures or referential integrity when I started using it. There are still limitations to MySQL. For example, for work use I don't want an RDBMS where pulling the plug on the server can cause data corruption. MSSQL and Oracle can recover from such a situation just fine, MySQL not so much.
But as always, horses for courses.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.