Our increasing reliance on today's high-speed computer networks has become not only a necessity of doing business but a way of life. Network connections are seemingly ubiquitous whether it involves our cellular phone, laptop, an ATM or credit-card transaction. Almost every aspect of our society relies in some way on computer data networks. When there is a breakdown in these networks disruption is caused throughout many areas of our society, often affecting millions of people simultaneously.
Because of the complexity of services provided by these computer networks, it is becoming increasingly difficult to detect and correlate isolated occurrences of these disruptions. To provide proactive management of these networks the vendors of the machines running on these networks have often included software called agents. These agents monitor the health of the individual machine, and store statistics about the machine, related machines and/or services. These statistics are stored in areas of memory called the MIB (Management Information Base), and can be accessed by a network protocol called SNMP (Simple Network Management Protocol). Types of statistics monitored and stored can include but are not limited to interface utilization, cpu utilization, temperature, fan speed, power supply status, memory utilization, etc. There are literally hundreds of thousands of metrics that can currently be measured by SNMP. ByteSphere currently has one of the world's largest online MIB databases, comprised of both standard and vendor MIBs. This MIB database has been made public for all those involved with Network Management, as it is often difficult to find needed MIBs in order to manage a portion of the network.
Our flagship IT management product, OidView, is a platform called a Network Manager; it uses SNMP and enables the user to administer, diagnose and troubleshoot, and even monitor the MIB values in those agents. OidView can be used by anyone involved in the management of a computer or a device on a computer network. OidView is a modular Network Manager that offers several technology plug-ins, therefore allowing customers to expand and grow their arsenal of diagnostic, administrative, testing, and monitoring tools while building on their initial investment.
Performance Management uses a number of collection techniques to determine network, device, and resource utilization as well as a variety of other critical metrics; coupled with predictive modeling algorithms collected data can be used to determine MTBF, leveraged for capacity planning, or exported and shared with 3rd-party applications and systems for a wide range of functions. Jaguar’s Monitor Engine (ME) can be used by itself or integrated with any 3rd-party NMS; allowing for efficient data collection and delivery to a variety of data and reporting sources. ByteSphere has also solved the “scale” issue, and our design allows for efficiently polling millions of network elements and utilizes a very small footprint.
Application Management is crucial to determine the health and availability of your mission critical applications and websites. ByteSphere Agent, our desktop prototype, can watch protocol utilization on any machine on a particular segment, and trace network traffic. Jaguar’s App Management Engine (AME) will perform passive collection of application traffic and allow alerting and reporting of application related events and detect attacks when they occur. The AME can be standalone or will easily integrate with 3rd-party network management and reporting packages.
Inventory Management uses a powerful, extensible discovery engine to get a holistic view of the agents, devices, and applications on the network, how they communicate, interact, and depend on one another. Once the topology of the network and the applications that run on it is defined, network maps and diagrams can be automatically created from the discovered information and overall dependency information can be determined. This in turn helps with Event Correlation.
Event Correlation determines how disparate events are related and makes predictions and alerts based on pre-defined heuristics. These heuristics can be created and extended by the user to allow for ultimate control of event reporting in the IT infrastructure. The JaguarSX real-time Event Correlation Engine (ECE) will be performing these tasks.
Whereas diagnostic information can be retrieved directly by proactively querying the information from the MIB using SNMP, it can also be acquired simply by listening for messages sent from the agents in the form of SNMP Traps, Informs, and Notifications. These messages are also known as alarms. The agents that have the capability to send diagnostic information in these formats have to be told where to send the information. ByteSphere's Fault Management solution includes OidView's Trap Management module. It can act as a listening Network Manager and can be configured with comprehensive recognition, filtering, and notification mechanisms.
Fault Management is a critical part of network monitoring. It gives the user a central place to receive, acknowledge, and organize all important notifications from devices on the network. More sophisticated fault management systems have the capability of modifying the information in some way and using something called a Notifier to forward or send updated messages to both machines and individuals about these occurrences.
Filtering systems can be used to identify certain alarms and execute pre-defined actions or notification mechanisms. Using filters, Fault Managers often categorize alarms into named buckets, allowing managers to more easily administer alarms. Most Fault Management software also assigns known alarms a severity - which should indicate to the user how important that alarm is and how quickly it should be addressed. Severity levels often include:
info (Informational alarm - no action necessary)
warning (Take note of this?)
minor (This could be important.)
major (This IS important!)
critical (Take action ASAP!)