AppSense Performance Manager

The Smart Scheduler allows the intelligent management of CPU resource by mapping business priorities or SLA agreements to how the CPU resource is allocated. The Smart Scheduler algorithm provides real-time calculations of the correct level of resources required for each application. Integrating directly with the Windows Scheduler, AppSense Performance Manager dynamically adjusts the base priority of each task to ensure that the correct level of resource is always given.

Unlike the Windows Scheduler alone, where the application or user which 'shouts the loudest, gets the most', the AppSense Performance Manager out of the box configuration guarantees each user and application an equal or fair allocation of resources. Administrative accounts and system processes are granted an elevated share of CPU time, ensuring that administrators can efficiently perform their duties regardless of how busy the machine is.

Shares of the available CPU resources are assigned to users/groups and/or applications using Share Factors. If an application is assigned a share factor that is twice that of a second application, then the former will generally be entitled to consume twice as much CPU time as the latter.

CPU reservations can be used to provide mission critical applications with a guaranteed minimum amount of CPU time. For example, if an application is allocated a reservation of 20% it will continue to receive priority access to the processor resources while it is using 20% or less of the processor resources. After exceeding 20%, additional resource requirements will be addressed according to the assigned Share Factors. If less than 20% is used, the remainder is made available to other applications.

A Soft Limit is a mechanism of restricting CPU consumption by a given task when the CPU resource is required by a more critical application or user. Until the specified percentage limit is reached, allocation of resource is divided according to the CPU Share Factor. Additional CPU time above that specified in the Soft Limit is allocated the lowest possible priority. This ensures that the task is only permitted to consume more that its specified limit when the CPU time is not required by other applications, users or groups.

A Hard Limit restricts CPU consumption of a given task. Unlike Soft Limits this is a physical restriction to the resource received by an application. For example, an application with a Hard Limit of 20% cannot possibly consume more than 20% of the CPU time at any time.

Affinity is a mechanism of restricting the access to particular processors by an application or user. On multiprocessor systems, this can be used to minimize the context switching of a given application.

Thread Throttling™ allows the setting of system wide CPU clamping policies, which automatically trigger when the system is heavily loaded and force any 'over-consuming' threads within each process to decrease their CPU usage.

By automatically analyzing and optimizing the Dynamic Link Libraries (DLLs) loaded by applications, virtual memory overheads and system paging can be significantly reduced. Optimized DLLs are stored in a separate cache and loaded dynamically, leaving the original applications intact.

AppSense Performance Manager gives you control over physical memory, letting you automatically trim working sets based on application events and states, such as application startup, idle, minimized and in the background. This releases physical memory back to the operating system for redistribution.

User memory limits can be applied to restrict the amount of virtual memory utilized. Users can be warned and then prevented from launching additional applications when virtual memory utilization reaches critical levels. Application memory limits may be applied to individual applications, which can force them to close down gracefully when they consume excessive memory.

Network Throughput Limits and Quotas can be applied based on application, user/group, network protocol, port number and/or direction. Network throughput limits allow network bandwidth usage to be restricted. Network Quotas can be applied to monitor, block or restrict network activity after a set usage quota has been reached.

Policies are defined for processor, memory and network management by creating application groups and using them to define resource sharing, reservations and limits. Policies may be further restricted to specific user and group accounts. Application states may also be included to provide precise control over applications delivered to desktops and Terminal Services/Citrix clients.

The integrated enterprise software deployment architecture allows AppSense Performance Manager software and configurations to be managed centrally, so administrators can apply rules to all servers and workstations from one MMC Console. Configurations are created as standard MSI packages, meaning they can alternatively be deployed using third party deployment tools.

Users and applications continue to work throughout any configuration changes as new settings are applied dynamically without the need for a reboot. The integrated Auditing framework records important security and performance events in industry standard formats such as system event log, e-mail and SNMP, or integrated into MOM using a free MOM Management Pack.

AppSense Performance Manager Management Reports provide the information necessary to record the use of system resources on a machine running AppSense Performance Manager. This can be used to charge for the use of a machine according to the workload. In addition this can also be used for Capacity Planning purposes.

Reports can be obtained concerning the overall use of CPU and memory, in addition to detailed reports on the consumption of resources by given applications or users