Top 10 reasons why you should KEEP your
legacy network monitoring system
Humorous.... but quite true.
- Ping and interface monitoring covers all your needs.
No need to bother with those pesky infrastructure metrics.
- Who needs those modern micro-second response time measurements.
Good old millisecond resolution is enough for your needs.
- Link utilisation graphs show that 100,000% usage. Must be some type
of new statistical metric.
How can a link be more that 100% ?
- Wireless monitoring... don't need to monitor futuristic technology
like that yet.
- Great vendor MIB support... except for the equipment you have.
- IPv6 support (ratified in 1998)... you'll never need that stuff.
- SNMPv3 (ratified in 1998)... who needs authenticated/encrypted v3 traps anyway.
- Five minute polling of packets/errors/discards/broadcast/multicast is good enough.
Who really needs 1 minute resolution to track down those multicast storms ?
- Custom OID monitoring works like a treat. Just can't get enough of it.
One day it might even monitor the stuff you've been asking for the last five years too!
- Operating system security updates - who needs that hassle.
Once a year is enough to keep your security team happy.
- Cheap to deploy - runs in a VM with dedicated attached disks.
Who uses those new technologies Hyperthreading/RAID/SAN/NAS anyway!
Oops, that's 11, but here's a few more ...
- Dashboards with heat maps, stacked bar graphs and scattered pie graphs.
Eye candy use only. Any competent Statistician will recommend never use
them because they are misleading and useless - basic 101 of statistical analysis
and presentation.
- Informative technical blogs and ebooks cover all your unanswered questions.
It's all in the buzz words - CAPEX, OPEX, OT, TCO, IoT, IIoT, BYOD, Smart Data,
RESTful API, Dashlets...
- Smart Data - must be all those extra bits in those bytes.
- One liner release notes with no history of fixes and enhancements.
Must be the new way of doing structured release engineering ?
- Flexible ratcheted pricing model allows you to only pay for what you monitor.
Who's got the budget to monitor their entire network anyway.
- Five annual technical support cases... who'd need more than that in a year.
- Customisable dashboards... just can't have enough clocks, logos and
pie graphs to fill that empty screen space on your 30 inch monitor!
- One second diagnostics poller is really useful... even if router and switch
vendors only update their SNMP counters every 10 seconds.
Hmmm! maybe that's why those graphs peak at 1000% for a second and then 0% for nine
seconds, and without high precision timers on the agent data collection side, the graphed
values are misleading.
- Read Only RESTful API... now that's damn useful! and there is always a spare resource
sitting around idle who knows how to program in Perl/Python/HTTP/CGI/JSON/etc.
- Next release is coming really soon.... er or later.
Can't wait to install version 6.0.1 with the latest hotfixes!
