Key Features of a Reliable IP Monitoring Tool for z/OS

Network monitoring tools are helpful in monitoring and improving the performance of network services. However, your network monitor must not hinder the performance of the network by consuming any substantial amount of bandwidth. For z/OS and TCP/IP, you will require a tool that doesn’t use techniques like SNMP (Simple Network Management Protocol) and constant packet tracing to provide real-time monitoring reports, as these require more than minimal amounts of bandwidth and CPU resources.

Importance of a potential tool
IP has become very crucial for many organizations with z/OS installations. Without proper monitoring of the IP stack, it is not possible to monitor IP performance in real-time. A monitoring tool is required to collect data about network operations in real-time and alert administrators when even minor variations occur, as sometimes ignoring even minor issues may later result in severe problems.

Resource uptime monitoring
Monitoring the availability of resources ensures that the services are operating normally. Apart from monitoring the IP stack just for up-time or down-time, the monitoring tool should also keep track of abnormal activity on z/OS systems, like low connection counts or high packet rates.

Monitoring performance of IP stack
Performance monitoring is typically the real-time monitoring of the IP stack. The metrics for defining IP network performance are not as clear as those for Systems Network Architecture (SNA). Apart from obvious network faults, the IP monitoring tool should also report ongoing network issues regarding fragmentation, retransmissions, routing errors, etc. It should also alert administrators to serious issues like resources reaching their maximum operational limits, etc.

Monitoring network activity
An IP firewall blocking inappropriate access to z/OS systems doesn’t necessarily mean complete protection, as it will not stop security breaches that occur within an internal network. Though network security is not the chief role of an IP monitoring tool, a monitor does provide logs with respect to failed and rejected connections occurring within the network. It also alerts administrators when performance monitoring finds unusually high activity, which can indicate a denial-of-service attack.

Efficient Monitoring
The frequency at which the monitored data is collected plays a crucial role in IP monitoring. If the frequency of data collection is within seconds, it may consume more bandwidth and other resources, impacting network performance. At the same time, if the data-collection frequency is in minutes, the credibility of the data becomes questionable as it is not real-time in nature. The tool should have a built-in or user-defined frequency for collection of data, thus enabling a good balance between real-time data and utilization of resources.

Many organizations use SNMP to access IP-related data. But z/OS does not collect SNMP by default. SNMP has to be configured and then activated. Though SNMP provides detailed z/OS IP information, it uses distributed agents and has to poll daemons constantly with multiple GET requests for collection of data. That may require significant overhead in the network. With the amount of bandwidth it requires, real-time monitoring with SNMP is out of question.

Tracing every packet is also an inefficient way to monitor IP networks. Analyzing every packet as it passes through the stacks makes considerable demands on CPU time, paging rates, and storage requirements. And as the amount of IP traffic increases on the network, so do the monitor’s demands on the system. Just when you need your monitor to run most efficiently, it draws more resources to itself.

Extraction and presentation of data
The presentation of complex data in an easily readable format is also crucial. Presenting data in a browser-based GUI enables monitoring of multiple LPARs, stacks, interfaces, routers, buffers, TCP and SNA applications in a single screen. As the GUI presents data in an easy-to-read format, even less technically sophisticated staff, like help desk workers, will be able to understand. Thus it relieves the experienced mainframe administrators from monitoring tasks, so that they have time to focus on more important work.

An ideal monitoring tool not only provides real-time reports and immediate alerts, but also runs efficiently within available network resources. Constant packet tracing and SNMP are not ideal methods of monitoring. We should also expect the tool to be making all the z/OS operations easier to navigate through and read without consuming valuable network or human resources.