At one of our marinas the Cisco DPC3939B Gateway is slow to respond when we have about 45 devices at peak usage, around 8-10 PM each evening and especially on weekends. Boaters complain about web pages not loading and streaming video (Netflix and the like) constantly buffering. Devices fail to attach via WiFi, probably because they time out before receiving a DHCP address from the gateway.
The level of service is Comcast Business Deluxe 100 Plus 100/20 Mbps. The network consists of the Cisco DPC3939B Gateway and 3 Engenius ENH700EXT access points connected with cat6 and dumb gigabit switches. No other router or DHCP server on the network. Another of our marinas, using a Comcast SMCD3G gateway with much the same network configuration and same level of Comcast service, has none of these issues.
A PC running PRTG Network monitor attached directly to any one of the DPC3939B ethernet ports shows average ping times to the gateway of about six ms, but max ping of hundreds of ms, with lots of variability (jitter). When this occurs a pc running PRTG Network monitor polling the dock's access points shows we are using about 30 mbps of our theoretical 100. Speedtest.comcast.net on a pc connected directly to the gateway starts at 75 mbps but often hangs near the end of the download speed test and gradually crawls to a stop.
Resetting the gateway to factory default does not change the symptoms. Pinging the gateway with only a pc directly attached after a reset shows an average of 2ms but max of 15 to 20ms.
When Comcast first-level support attempts to schedule replacement of the gateway I am told that there is a problem with the way our account is set up in their system and the gateway replacement cannot be scheduled. First-level support has twice escalated to Comcast scheduling. To date Comcast scheduling has not contacted me as promised.
Comcast first tier support has finally scheduled a replacement of the modem. Apparently there were "too many outlets", which created an error in Comcast's admin system. To their credit, each of four tech's attempted to escalate to scheduling but the system defeated those attempts.
To Comcast's shame there appeared no effective way to escalate the issue other than my calling repeatedly and working with each tier 1 tech, only to find out that no replacement had been scheduled. Four hours of talk time seems extravagent to get a truck to roll.
Comcast replaced Cisco DPC3939B with Netgear CG3000DCR.
Ping times to the gateway from a pc directly connected via ethernet are around 1ms. Web pages load without interruption and video is not buffering. This is under a light load of 26 devices and 3 Mbps but it is encouraging.
This morning the PRTG network monitor announced that it had noticed an unusual condition and dryly commented that the router's average of 104 msec to load the admin web page is quite low. Well, yes, given that the Cisco would often require 2,500 ms to load, this is quite unusual.
Max ping times during the Friday night rush hour were below 10 ms and averaged 1ms. So far, the problem appears to have been resolved. If the Netgear handles the load throughout the weekend, I will log the ticket as closed.