More-VRP works on the operating system's inter-process level, underneath the database. It is fully transparent to the database, so that no architectural changes are required.
More-VRP technology 'listens' to all the traffic between the database and the operating system, without interfering, correlates the data with information gathered from the database, and when a rule is triggered, More-VRP instructs the operating system to allocate the selected amount of CPU and IO to a specific operation. This is achieved by creating narrower virtual resource lanes to the CPU and IO drivers, and redirecting managed transactions to these virtual lanes. More-VRP monitors all of the various processes, and generates detailed operational, statistic & analyzing reports.
More-VRP can control any SQL (Select statements), DML (Inserts, Updates, Deletes and Merges) and even DDL (Alter, analyze, Rebuild etc.)
Instead of stopping (Killing) heavy resource consumers, or re-scheduling them, you just set the amount of CPU power and IO bandwidth you want to allocate to them and let them run in the background without interfering with the online work.
MORE-VRP supports:
Operating systems, including: Red Hat Linux, SUSE, Sun Solaris, HPUX, IBM AIX, Windows, and Tru64.
High Availability
In enterprise organizations where maximum availability is required for business continuity, an availability solution for MORE is absolutely crucial. MORE-VRP can be configured in a clustered environment to ensure maximum availability for load management, and during server failures.
Footprint
MORE-VRP agent footprint in an average database is 0.5% of a single CPU, where, for example, a machine with four CPUs, MORE-VRP uses ~0.12% of the total machine. The system's engine is installed on a Windows PC/Server on the network, and collects various statistics, correlates them, then logs them into an internal repository.
Monitoring
The system can be easily monitored and controlled by pre-defined rules, and corrected as required. When a rule is defined, the virtual partitioning of the server's resources is performed immediately. Fully transparent to the database, it does not require special configuration at the DB level. Data is clearly displayed on MORE's unique system dashboard. The dashboard allows users (DBAs and managers) to identify problems in the SLA, and pinpoint any heavy transaction that may be the cause of the problem. Users can create rules and prioritize the resource usage directly from the dashboard.
API
MORE-VRP easily integrates with various monitoring tools by means of a built-in API. Any 3rd party monitoring tool such as Tivoli, Unicenter, HPOV, Grid control, Patrol, and others, can connect More-VRP using the built-in API and read MoreVRP counters, or alternatively, activate the VRP management mechanism.
Reports & Alerts
MORE-VRP is equipped with built-in state-of-the-art reporting capabilities, To best evaluate database behavior and enable users to create efficient rules, MORE constantly logs the traffic into a repository. The repository can then be analyzed using MORE-VRP reporting capabilities, to drill down into problematic peak periods, SLA trends, and abnormal events. However, users can easily convert data to standard CSV/XML files and integrate the repository with any other 3rd party BI tools, including BO, Congnos, Discoverer and others.
Installation
Installation takes 2-4 hours without any downtime or changes to the application or database server. Full implementation which includes Performance recording, performance analysis (Using More Reporting capabilities) and the creation of automatic rules, takes 3-5 days in average. MoreVRP is easy to use, but if needed, More-IT-Resource provides a full online training service.
|
