Turnkey Engine
Turnkey Engine”
Back in the 90s a group of German engineers and the computer network initially distributed to over 100 computers that work on the first version of Linux. It was a great success and everyone calls it the dawning of a new technology that will change the computing world forever.
What these engineers did not understand was how the database engine has worked over time and how this really set the trends for the development of hardware. RDBMS Data were calling for massively parallel processing (MPP) systems which offer dozens of processors on a single server platform. Just when we thought the mainframe was dead – with computers grid, we have witnessed the birth of the mainframe for open systems.
To compensate for these systems MPP software architects created a layer of middleware to escape these monsters. Now we have GPP these systems in the center of the universe and all these servers, middleware dance around him.
For the next few years, companies filled out data centers with hundreds of servers and tens of middleware systems MPP. Cooling Supplies and power running at its peak and administrators data centers do not know how to support this additional hardware in the future.
It was the cradle of server virtualization. In years, server virtualization has become the main objective of any business who needs to save the infrastructure. Middleware server consolidation is an easy task, and a huge success with pride shared by those responsible project.
However, one area that has not been so successful in consolidating the material. MPP systems will not go away quietly. It turned out that the database systems were too difficult to control for the frenzy of server virtualization. Yes countless efforts have been made to move the databases of these monsters, but the virtual world could not provide the performance needed supporting databases.
Remember that our grid computing history? The vision of sharing a small army of computers to produce the same computing power and ground systems MPP seemed lost forever. Until the horrible truth about the MPP systems became apparent. Execution of large MPP systems not only are large consumers of energy, but maintenance costs are also expensive. Buying a replacement system the member was the only option. No. But spending money increasingly tight in recent years. All this has become a dead end 22. It must invest more to increase the costs!
providers databases to the rescue: database networking technology seems to be the way out from under the massive weight of these systems MPP. Take a couple of powerful servers inexpensive dual or quad core and distribution the load between multiple servers. Not only do you instantly get high availability, but said gains scalability beyond the limits physical platform of the MPP. There are two main methods in the pursuit of databases on the network, all to share, not common.
The whole category is dominated by shared Oracle and Sybase. Both systems are capable of immediate failover process server database participants Autumn, also known as. high availability. And both can dynamically scale your CPU resources by adding more servers on the network, and both systems can balance the load between all servers. Oracle RAC and Sybase ASE Cluster Edition is the most sophisticated system available today. If you want to maximize their existing hardware or are looking to replace a waste of energy and maintenance costs eating monsters MP, these two databases are the weapons of choice.
The second approach is not shared. Microsoft SQL Server 2008 Data Warehouse Federation represents the leader in this category. Contrary to the technology to share everything, nothing common approach has a clear distinction between local data and overall. The approach to data federation allows the combination of data stored locally on several individual databases. It acts as an aggregator of multiple databases data. It's not as sophisticated as the common goal of all, but does the job well.
A third man, Sybase IQ Multiplex. This system uses a hybrid approach, shared data, but no cache. This is very unique and no supplier database has something similar. Sybase IQ Sybase Data Store Engine. A column vector database which sets new performance records for reference a single node with different authors in the group and an almost unlimited nodes reader. But there is a warning: Never try to run an application on the OLTP system. This system is designed for data storage and reporting mass analysis, a perfect combination for the data hungry BI tools.
There is a database of technology providers offering some network database of some flavor. This article is not intended to create a business competition between all these systems, but a starting point for your imagination. The bottom line is that energy conservation becomes increasingly important and software vendors offer solutions to maximize the low cost equipment and low energy consumption. The future belongs to databases network.
About the author:
Peter Dobler, founder and owner of Dobler Consulting Inc, who has over 20 years of database management and data migration experience. He and his company are committed to provide turnkey data migration project management and data migration execution to delight its customers. Read more about Peter Dobler at http://www.peterdobler.com and http://www.doblerconsulting.com
