66 E L EC TR I C AL CONNEC T I ON
SUMME R 20 1 6
GETTING TO THE BOTTOMOF THINGS
C
omplex and sprawling networks
can be difficult to troubleshoot,
and network professionals
spend approximately 25% of their time
troubleshooting.
NETSCOUT director – APAC channels
Amit Rao says, “Finding the root cause
of network issues is time consuming. If
they’re intermittent issues, it can seem
almost impossible to find and resolve
them. However, by taking a methodical
approach, it’s possible to troubleshoot
enterprise network problems.”
1. INFRASTRUCTURE PERFORMANCE
End user complaints often signify
that there is an infrastructure issue.
However, when application servers and
infrastructure devices are operating
normally, obvious error states can’t be
located and legacy network monitoring
tools report ‘green’. Finding the root cause
can be challenging.
Possible causes include bad cabling,
network congestion, server network
adapter issues or DNS issues.
There are four steps to troubleshooting
these issues:
>
use existing monitoring tools
and extract information from
SYSLOG receivers.
>
check server and network device log files
to understand if there are connectivity
issues from the NIC side.
>
examineWAN links and logs to
understandwhether traffic-shaping
devices or policies are affecting
performance.
>
check errors including web server, load
balancer, and application log errors.
2. NETWORKSERVICES
There are numerous issues that can
affect network services, such as DHCP
issues or a slow DNS response. Possible
causes include misconfigured DHCP or
DNS servers, duplicate IP addresses
caused by overlapping DHCP scopes, rogue
DHCP servers, or users manually assigning
static IPs. This can enable a ‘man-in-the-
middle’ attack and create significant
security issues.
To troubleshoot, first confirm proper
configuration of authorised DHCP servers.
3. PROVE IT’SNOT THENETWORK
Most of the time, the network is not to
blame for performance issues.
People blame the network due to lack
of visibility into network operations, not
enough bandwidth, network complexity,
insufficient network expertise and lack
of effective troubleshooting tools.
To troubleshoot, you should use packet
captures, gather network data, review
dropped packets, check for excessive
retries and congestion in capture files. You
should also check network device logs and
ping to check response times, as well as
using tracert to verify that the network
path is correct.
4.WIFI ANDBYODTHREATS
WiFi networks, combined with bring
your own device (BYOD) policies, can
create security and performance issues if
not managed carefully. These can include
chatter, dropped connections, excessive
bandwidth issues, and poor device
behaviour from users (such as streaming
music), congestion. The sheer number of
devices can swamp the network.
To troubleshoot, conduct regular WiFi
SSID surveys to detect rogue access points
and routers. Look up MAC addresses to
discover the types of devices attached to
networks and implement MAC address
filtering if necessary. Also, understand
that some devices are well known for
causing problems if improperly configured,
for example, Apple TV Airplay can badly
impact performance.
5. POORWIFI PERFORMANCE
When the WiFi network is
underperforming, network teams should
check for frequency interference, rogue
routers (such as phones being used as
hotspots), misconfiguredWiFi routers, and
compatibility issues between certainWiFi
clients and routers. Even excessive heat
can cause strange symptoms.
To troubleshoot, teams should
regularly use an SSID scanner to
identify rogue routers and APs in
infrastructure, remember that strange
DHCP behaviour is an indicator of rogue
DHCP servers, relocate routers that
may be suffering interference due to
proximity to EMI sources, and ensure
that all WiFi devices are within their
designed operating environment.
6. INTERMITTENT PERFORMANCE
Transient issues can take time and,
sometimes, luck to capture, diagnose, and
resolve. Causes can include cabling issues,
external sources, power fluctuations,
hardware failures and excessive heat.
To troubleshoot, rule out logical
sources, then look for illogical sources of
interference. Track occurrences of the
specific performance issue and look for
patterns. As always, start at the physical
layer, using a cable tester to see if the issue
is related to cabling.
Amit says, “Understanding how to
troubleshoot the most common issues
can reduce the amount of time network
professionals spend on issue resolution.”
> NETSCOUT
www.netscout.comNETSCOUT
EXPLAINS HOW
TO CONQUER THE SIX MOST
COMMON ISSUES IN THE
ENTERPRISE NETWORK.
TROUBLESHOOTING