Most
global business enterprise use different business applications
installed on servers within the organization’s network or data centers
to provide services to it hosts. These applications on the server
necessitates constant monitoring for resource usage like disk space,
cache, CPU etc., and other issues that may affect the performance of the
server.
To
ensure an effective network monitoring, the monitoring design should
adopt a few basic principles, it should cover every aspect of the
organization like the network components, security, reporting, problem
detection, and resolution and network maintenance.
For
successful monitoring we need the data or information from network
components being monitored, a monitoring application to collect this
data for processing, presentation (Reporting) and performance alerts
(Alerts) and a protocol to transmit the information between the network
components being monitored and the monitoring software.
Alerts
and reports complement each other, alerts let the network admin know of
potential problems and reports provide data to identify the root cause
of the problem.
Networks
have baseline description of what is normal network operation
performance, Alerts are sent out when the values of this set parameter
deviates from this established baseline thus helping in early detection
of issues which in turn contributes to the smooth functioning
network/uptime.
Basic Network Monitoring Strategy
Availability
monitoring: Requires the monitoring of all IT infrastructure to ensure
their availability to perform the organization’s and user requirements
thus a 100% uptime of the services offered by the organization.
Availability monitoring enables the monitoring of the network devices to
ensure smooth operation, bandwidth availability to ensure delivery,
data space availability and monitoring system level services to ensure
enterprise critical applications remain functional. Technologies used in
availability monitoring include
•           PING: Tool used to test reachability and availability of host IP network.
•          
SIMPLE Network Management Protocol (SNMP): An information exchange
protocol used in a host network that includes networking monitoring
software that is widely used.
•           Syslog: A message logging system that enables devices to send event notification in IP networks.
•          
WMI: windows based systems may provide data to monitoring systems which
uses the data to create performance report of the servers, host
machines and network components. This vital data can be collected using
Performance counters, Windows Management instrumentation (WMI) and Event
Log
•           IPSLA: A feature of cisco that can measure availability of WAN links and their capacity
•           Telnet: Telnet is used to check for the availability of a device on a network when ping is blocked.
Interface
monitoring monitors the entry and exit points of packet on a device on
the network. Types of device interfaces include Fast Ethernet, Gigabit
Ethernet and high-speed fiber optics. These interfaces have to be
monitored for error, packet loss and utilization limits thus the
information garnered help identify issues that could cause poor
application/service performances.
Disk
monitoring is another strategy was to ensure proper disk management for
the maximum utilization, performance error check, file stats, I/O
performance etc. on the server. It enables the network administrator
plan for space and detection storage related problems.
 
Hardware
monitoring: The network infrastructure constitutes many hardware
devices like the route and switch, server computers etc. Losing the
function of one of these hardware like over-heating of a core switch
could result in a downtime so it is important to monitor their health
and performance too. Certain metrics like temperature, fan speed, power
supply and CPU utilization can be used to monitor hardware performance.
 
CLOUD
Networks
have evolved and become more complex in past few years, organizations
now have hybrid cloud environments and third-party contents so what
worked in a traditional client/server centralized and enterprise
controlled environment may not be effective in a decentralized
environment thus organization must reevaluate their network monitoring
strategy. The network administrators must determine if existing network
monitoring tools can provide them the require end-end visibility across
internal private network, external public cloud infrastructure and the
connectivity between them.
Traditional
networking monitoring tools work by collecting information through
Device polling, Active probing, Internet Protocol Flow Information
Export (IPFIX), Packet capture and Data logging and this is only
applicable on the infrastructure where the enterprise owns the network
devices. But with cloud infrastructure you do not have visibility into
how applications perform across the network. Infrastructure where the
enterprise owns the network devices. But with cloud infrastructure you
do not have visibility into how applications perform across the network.