 |
Network Management & Performance Monitoring with eG Enterprise
|
 |
|
|
Network Monitoring as Part of an End-to-End Monitoring Solution
Network monitoring is an integral part of eG Enterprise's end to end monitoring capabilities. eG Enterprise agents track network availability, network performance and usage for servers and network devices in a data center. These metrics are combined with the metrics, from in-depth server & application monitors to provide a complete end-to-end view of every layer of every tier of the infrastructure that supports a business service.
The network performance monitoring is done in an agentless manner - i.e., without requiring any software on the servers or network devices. SNMP capabilities supported by most network devices are used for remote network monitoring. To support the increasing security requirements of network administrators and at the same time, to support legacy devices, eG Enterprise includes support for all three versions of the SNMP protocol - v1, v2, and v3.
SNMP Monitoring using eG Enterprise
SNMP-based monitoring is used to track the performance and usage of a variety of network devices including switches, routers, hubs, printers, firewalls, web accelerators, application switches, SAN switches and other network devices. eG agents deployed in strategic locations can poll the different network devices to collect a variety of statistics. The metrics to be collected are determined based on domain expertise with each of the devices being monitored. The metrics collected from each device vary depending on the MIBs supported by the device.
Key Network Performance Metrics that eG Enterprise Monitors |
| Device boot time |
Device uptime (since last reboot) |
| Availability of each of the network interfaces of a router, server, switch, hub, etc. |
Bandwidth usage on each of the network interfaces |
| CRC errors on each interface |
Collisions on each network interface |
| Speed of each network interface |
Queue length for each network interface |
| Packet drops due to buffer overflows for each interface |
Network connectivity to each network device and server |
| Network latency to target devices and servers |
Hardware status of the network device including current temperature, voltage, etc. |
While SNMP polling is the primary mode of metric collection, eG Enterprise also supports SNMP traps. The devices can be configured to send traps to a specific eG agent. Upon reception of a trap, the agent analyzes its contents and decides if the management console needs to be informed of a problem with the device.
 |
Figure 1: Adding new SNMP OIDs for monitoring using
the eG SNMP MIB Browser |
| |
 |
| Figure 2: Monitoring each of the network interfaces of a Cisco router |
Standard SNMP MIBs including SNMP MIB-II as well as proprietary MIBs (e.g., Cisco) are supported by eG Enterprise. To allow administrators to monitor network devices that are not supported out of the box, or to add custom network monitoring capabilities, eG Enterprise includes a MIB browser using which administrators can load any SNMP MIB into the eG management console and pick object identifiers (OIDs) that need to be monitored. Extending the network monitoring capabilities of eG Enterprise is thus very simple. No additional programming or scripting is needed.
To detect network connectivity issues and slowdowns, eG Enterprise uses the ICMP protocol. Using ICMP, eG Enterprise tracks the availability, network delays, and packet loss during communication to the network devices and servers. In the event network issues are detected, eG Enterprise offers additional diagnosis by reporting hop-by-hop delays.
Administrators can use this information for better network management - by detecting the hop that is responsible for a network slowdown, administrators can quickly initiate the necessary corrective action.
eG Enterprise uses a scalable network managment & monitoring architecture. To monitor thousands of devices, the devices can be partitioned and assigned to different eG agents. Additional pollers (agents) can be added as the target infrastructure expands.
Network alerts are generated based on threshold violations. The thresholds are either pre-specified by the administrators or they are automatically determined based on past history of the metrics. Such automatic thresholds are ideal for usage metrics, such as network traffic to and from each interface of a router.
A key network performance management feature is the ability to auto-discover network topologies and to automatically correlate between the performance of networks, servers, and applications. For instance, if a core router goes down and all the servers connected to it become unreachable, eG Enterprise auto-discovers the connectivity between the routers and the servers and correlates the server and network alerts. In this case, the server alerts are downgraded in priority to indicate that the network alert is the root-cause of the problem. Thus, eG Enterprise's automatic correlation capability allows administrators to be informed of the root-cause of problems so they can focus their attention on the cause of problems and not be distracted by their effects.
Licensing of the Network Management and Monitoring Software
Network monitoring is an integrated part of the eG Enterprise suite and is not separately licensed. Furthermore, eG Enterprise does not price its solution based on the number of network devices monitored. Additional agents need to be procured and deployed as the size of the network infrastructure to be managed increases.
Key Benefits of the eG Enterprise Network Monitoring Solution |
| Real-time network management and alerting on network performance issues so administrators are aware of problems as soon as they happen. |
| Automatic correlation of network performance with server and application performance, so administrators can quickly determine where the bottleneck lies. |
| Automatic baselining of network performance so administrators can be alerted proactively whenever network usage exceeds historical norms. By acting on these early warning indicators, administrators can avoid catastrophic failures in the future. |
| Scalable network monitoring architecture that allows additional agents to be deployed for SNMP polling and trap handling as the scale of the infrastructure being monitored increases. |
| Licensing of the monitoring solution is not tied to the type of devices or the number of devices monitored. |
|
|
| |
|
Get Your Free Trial
