![]() |
![]() |
|
For decades, monitoring solutions employing the manager/agent model have differentiated themselves with the intelligence and analysis capabilities incorporated into the manager component. The software agents that collect and report measurements pertaining to the health of an IT infrastructure have been viewed as commodity items that perform similar tasks. This line of thought undermines the key role that agents play in a monitoring solution:
The eG Single Agent Architecture for Network, Server, and Application ManagementThe eG agent architecture has been designed to benefit both small and large organizations. The key features of this architecture are: Unique Single Agent technologyThe eG Enterprise suite requires just one agent to manage the operating system and all the applications executing on a server. The eG agent uses various native interfaces to monitor and manage over eighty different applications and network devices. This unique single agent licensing policy ensures that administrators no longer have to worry about which applications are deployed on a server and whether there are sufficient licenses and appropriate agent monitoring modules available to manage these applications.
Low overhead through multi-level monitoringTo minimize the overheads involved in on-going monitoring, eG agents support multi-level monitoring. On an on-going basis, the most critical metrics are collected. When a performance abnormality is detected, the agents can make more detailed measurements. This multi-level monitoring capability ensures that the overhead of an eG agent is minimal – typically, 0.1-0.3% CPU Fast deployment through 100% web architectureDeploying the eG solution is a breeze. The agents communicate with the manager using web protocols – HTTP or HTTPS. This ensures that administrators do not have to bother about proprietory protocols being used in their networks. Furthermore, since all communications are initiated by the agents, firewall rules need not be changed in order to get the eG Enterprise suite operational. The agent initiated communication model also facilitates remote monitoring across geographic boundaries. This architecture is particularly ideal for managed service provider (MSP) environments, wherein the agents may be located within different customer Intranets, and the manager may be in a central server farm. The use of HTTPS ensures that it is no longer necessary for IT administrators to set up expensive virtual private networks between a network operations center and the infrastructure being managed.
Zero maintenance through automatic upgrade capabilityOne of the long-standing drawbacks of agents has been the need for on-going maintenance. The eG architecture effectively addresses these limitations. New functions to monitor new applications can be dynamically downloaded to the agents. Product enhancements are also delivered as upgrades that the agent can retrieve periodically or on-demand. Administrators can centrally control when and which agents are enabled for automatic upgrades using a web browser interface thereby facilitating zero maintenance during operation. Cost-effective, hardware independent licensingThe eG license controls how many agents are simultaneously reporting to the manager, not where the agents are deployed. Hence, administrators have complete flexibility in determining when and where the agents need to be deployed to enable comprehensive monitoring of the infrastructure. Moreover, the agent licensing is independent of the operating system of the target servers offering administrators complete flexibility in using their agent licenses and ensures easier migration across operating system platforms. Flexible yet rapid integrationWhile SNMP is a standard way of interfacing with network devices, there is no real standard for monitoring systems or applications. Different operating systems offer different interfaces for measurement. Each application is different in its functionality and in the APIs that it supports to expose performance information. To handle these varied infrastructure components, eG agents adopt an open architecture. The eG agent uses various approaches to interface with applications – e.g., it supports SNMP for network devices, Java Management Extensions (JMX) for some middleware servers, proprietary APIs for different web servers, SQL for interfacing with databases, perfmon and Windows Management Instrumentation (WMI) for Microsoft applications. The Integration Console plugin allows developers to add new capabilities rapidly to the eG agent in a few hours, not over days or months! Multi-perspective monitoringA single eG agent is capable of performing different types of functions. Taking an external perspective, an agent can simulate user requests to an application (e.g., web server, database server, DNS server). The same agent is also capable of using the application APIs to collect internal metrics of server processing time, queue lengths, error rates, etc. Multiple agents can also be deployed to assess the performance of an application from different locations. The flexibility it offers to administrators makes the eG Enterprise suite an attractive choice for infrastructure administrators.
|
||||||||||||||||||||||||||||