I recently had a similar situation, and the issue turned out to be resolved by upgrading from a DOCSIS2.0 to DOCSIS3.0 compliant cable modem. A similar situation also occurred when I upgraded from DOCSIS1.x to DOCSIS2.x some years ago.
Not saying that's what the issue is, but it's something to check especially if you can confirm your ISP changed/updated DOCSIS version or sub-version support between the time your service was working "OK" and the time it started "Failing". An ISP also may not support ALL the standards associated with a given DOCSIS version, including "downward compatibility" and the same is true regarding individual modems, which further complicates the issue.
In my case, the ISP was providing consistent service as far as THEIR
tests on THEIR
end showed...but at my end it turned out to be the cable modem (Linksys/Cisco CM100) that proved to be the "internet" bottleneck (aka performance and
availability inhibitor). It caused both repeated and and/or intermittent internet performance "hiccups". I first noticed symptoms when YouTube buffering started "stalling" and video would stop playing for awhile...then resume after minutes
passed...but my thought processes didn't immediately make the connection. I just put it down to recent changes in YouTube download functionality (I think driven by an anti-copy/download scheme, they stopped using temporary files on the system drive and started using only limited real-time RAM buffering (which causes other problems...bu that's another issue entirely)).
The ISP support team's standard "diagnostic" flow-charts did NOT make the diagnosis and did not help resolve the issue. It was left to me to figure out...both years ago and just recently...very
disappointing since I would think this would be a common problem whenever
the ISP changes their standards compliance at any level.
BTW, while in both cases the ISP still claimed
to "support" the prior DOCSIS version and
the old modem, performance at my end was still severely affected in each case.
In the most recent instance, when I changed to the newer DOCSIS supporting modem, ISP "download" performance increased from about
above 30Mbps (near the ISP service tier
maximum for "download") and "upload" increased from about
5Mbps to consistently
about 20Mbps (also near the ISP service tier
maximum for "upload") using speedtest.net
. It did also depend somewhat on which ISP target location I used to run the test (e.g. a San Diego server v. a San Francisco or Austin server). In addition, server distance, capabilities, and real-time loading (other users) can affect reported performance at any given time.
The new modem, a Surfboard SB6121 also provided "channel bonding", which when also
supported by the ISP could theoretically also improve performance and consistency. For more info on that, see internet search terms [channel bonding in cable modems]
Also, keep in mind that simply changing from a router that supports "10/100mbps" to "10/100/1000mbps" (aka gigabit) will not likely address things if DOCSIS is the issue unless the performance tier guaranteed by what you're paying for your ISP is very high (eg greater than 100Mbps)...although you might see a computer-to-computer performance improvement on your
side of the router assuming your computers also each supported "gigabit" in both hardware and software. Wireless performance would be separately limited by your wireless connectivity even if all devices on "your" side claimed to support "gigabit" (e.g. just claiming
gigabit network capability
doesn't magically make it so in practice).
* Complicating the issue...
Terminology Note: when discussing network bandwidths there is generally
speaking, no difference between sites that use the label Mb
ps and those that use the label mb
ps when evaluating network performance. In these
" is the same as "b
" and are both referring to mega-"bits
" per second. However re: "M" v. "m", I think the sites using "M" instead of "m" are simply trying to strictly differentiate between the standard conventions of "M"=mega (10E+6, million) and "m"=milli (10E-3, 1000ths), although "m" is often used interchangeably when discussing network metrics.
There is however, a huge difference between MB
ps (Bytes per second) and Mb
ps (bits per second). Even so, YOU
have to be sure what is being referenced to be certain you are comparing apples to apples. Some test sites and/or reference materials do not properly distinguish between the two and just confuse the issue.
To make clear that I had a similar problem recently
and figured that nathan (and others) could be going through the same/simiar issue because of the timing. My ISP (without warning) changed DOCSIS compliance specifics. Even though the ISP "still supports" DOCSIS2.x and
the specific modem (CM100) I was using, I also had intermittent loss of and internet performance issues until I replaced it with a DOCSIS3.0 compliant modem.