I am dealing with a home network which has recently experienced excessive dropped connections. Connection to ISP is via dedicated 2.4 Gb microwave radio card mounted on house roof. The home network uses a Linksys E4200 wireless router. On the home side of the router are typically 3 wired devices and 4 wireless devices.
One of the wired connections is an HP laptop running VPN on Vista, the second wired connection is a desktop computer using normal applications (no VPN) on Windows 7. The third wired connection is an HP network printer.
Of the wireless devices, one is an HP laptop running Vista and no VPN, the others are handhelds accessing the internet.
When the connection is dropped, most devices loose internet. The wired laptop running VPN may continue without losing the VPN connection. I am initially unable to connect to an internet domain using a URL and I am initially unable to ping my ISP’s DNS server IPs. If I release the router’s IP and then renew the routers IP, the internet connections usually return immediately. The ISP support tells me my roof mounted radio card is not losing connection when these instances occur.
I have viewed the Event System log on the wired desktop. There were many log records showing Event ID 1014 and description reads, “Name resolution for the name some-domain.com timed out after none of the configured DNS servers responded”. A quick census of the log file showed great variation in the number of these warnings week to week. I plotted a histogram of these events and found the following.
Frequency
Weeks of ID 1014
01 – 04 51
05 – 08 36
09 – 13 31
14 – 17 3
18 – 22 7
23 – 26 8
27 – 31 3
32 – 35 0
36 – 40 0
41 – 44 0
45 – 49 0
50 – 53 6
54 – 58 16
59 – 62 77
I also started logging network activity using netmon.exe in hopes that I can see some change in activity preceding future occurrences.
Were my actions sufficient thus far?
Should I do something else to gather evidence to determine root cause?
Are there network changes that can be made to reduce these DNS events?