[imitchellg5]

Comcast came out last night and did something with my cable box. Afterwards, my Bravia wouldn't tune into any channels, so I had to reprogram it. No big deal. Then I realize that my internet is incredibly slow. So overnight I decided to leave my cable modem and Time Capsule unplugged to reset the modem, thinking that some settings or something had changed. Today it's still painfully slow:


It used to be:



Any ideas on what I can do, or should I just call and probably not get the issue resolved?

[reader50]

Call. Do some screaming. Don't threaten their dog on the first call.

[olePigeon]

Calling doesn't do anything. They'll just tell you it's your fault.

Also, it's XFinity now. They listened to their customers... so they changed their name to avoid all the complaints.

[imitchellg5]

Well that was easy and also a bit suspicious... I called, was put on hold, in the middle of hold, the speed surged back... but it's still not to the level that it was. When the lady came back, she said it was all fixed... could they have put a limit on me?

Now it's:

[reader50]

It might be a high-traffic time. Check again at a few other times of day, see if the speeds return to normal.

[mduell]

The high ping supports the congestion theory.

[Snow-i]

Sounds like they could have reprovisioned your modem to an economy bootfile. This wouldn't have changed anything on your billing account, but call and ask them to run "APT." They should know what that means.

Also, they could have been working in your area on whats called a "plant fault." This means there are signal related problems in the area that may or may not have affected you, which the average comcast phone tech would be blissfully unaware of. Whatever work they were doing may have expanded the problem to include you on the fiber node, which the tech that came into your house was there to verify/test. Unfortunately, you have no way to tell other than taking signal readings yourself. For example, a node 3 miles away went down, and to restore service until the node is fixed/replaced for those users, they diverted that node's customers to yours, effectively doubling (or more) its workload.

It could also be that the tech in your house was checking splitters on the outside, or inside of your house which may or may not be powered splitters. If you have powered splitters and they lost power, well, your connection would become extremely noisy and suffer from poor speeds.

The other alternative is that they did a "maintenence hit" on your cable box. This resets the box to its factory settings and would cause the box to have to download a bajillion updates, thereby hogging your bandwidth. You'd be able to tell if every couple minutes your box displays "DL" on its front display. If it has no display, you have no idea how to tell other than it all the sudden just speeds up.

Remotely, aside from what I mentioned, no first level tech or even second level tech would have any means of crippling your connection aside from changing your billing codes and reprovisioning your modem.

If you have an eMTA (Digital voice modem), then the first outcome I outlined is the most probable.

If not, I would place money on the plant fault scenario or the box-update theory or some combination of the two (they often coincide). The plant faults can take them days to weeks to fix, depending on the scope and severity of the issue.


To save you aggravation, give it a couple days and see if it improves.

[Snow-i]

Originally Posted by mduell 

The high ping supports the congestion theory.

Unfortunately, this is incorrect.

Lots of/Too many users on the fiber node would not cause higher latency, just reduced bandwidth. Higher pings usually describe poor signal levels either locally in the house or for wider area such as your entire fiber node (anywhere from 150 to 800 residences).

*Edited to add: Yes, in theory, too many users can cause spikes in latency. In practice with Comcast's fiber and coax network, that doesn't happen.

[imitchellg5]

Interesting. It's been running at full speed since the 23rd. I've never experienced any slowdowns during any time of the day. I admit I know very little about networking, but the "maintenance hit" idea seems like that may be what happened. Although it still seems odd that speed went right back up after I was on the phone with them? Can these updates be controlled from a remote center? Perhaps then the person on the other end stopped the updates? I have no clue what the cable box looks like, since it's in a box that's filled with black widows for half the year and I'm much too chicken to open it up

[Snow-i]

Originally Posted by imitchellg5 

Interesting. It's been running at full speed since the 23rd. I've never experienced any slowdowns during any time of the day. I admit I know very little about networking, but the "maintenance hit" idea seems like that may be what happened. Although it still seems odd that speed went right back up after I was on the phone with them? Can these updates be controlled from a remote center? Perhaps then the person on the other end stopped the updates? I have no clue what the cable box looks like, since it's in a box that's filled with black widows for half the year and I'm much too chicken to open it up

The person on the other end has no way to control the updates directly, however, can send a regular "hit" (like when they say they're sending a signal to the box) which basically sets the correct authorizations (known internally as tier masks) on the box. This, on newer DCH, DCX, and other m-card boxes will pause the updates until the reauthorization of the box is complete. If you were complaining about internet speed, this seems less likely.

In my experience, it sounds like you were in an area that had a plant fault which affected your entire area. Note that some users would be affected more due to local signal levels and sensitivities. Many users wouldn't notice a problem at all and for some it would be unusable. This is why they don't assign outage status nor do they assign top priority, but instead call it a plant fault or priority plant fault, though they do tend to get those issues fixed pretty quick (within a week). Unless the person you're speaking with has access to a tool called watchtower, which is used by the LMC (outage monitoring), they would have no way of telling other than signal readings from the modem which probably were not out of operational spec, just near the threshold. Even their proximity check tool on the standard troubleshooting package (known as grandslam) doesn't recognize these faults, and that is by design. All of the advanced troubleshooting tools were taken from the tier 1 and 1.5 techs that you are able to speak to.

To give you some perspective, the DC/MD/NOVA region has between 25-50 localized plant faults at any one time, and usually between 0-5 node outages (total service failure across hundreds or even thousands of users). Comcast has fiber and coax technicians fixing these network issues 24/7, and does little to give first and second line techs the knowledge of this advanced networking and how it affects the customer. First and Second line techs get paid very little (the going rate here is 12.02 an hour) and on average, have no technical experience other than Comcast's training courses which are, at best, obsolete. Their call-center turnover is almost 60% in 1.5 years, and have gone with a business model that puts little technical talent on the line with you but focuses on customer service. They don't invest in their people at that level, and just want warm bodies that can handle 95% of the calls for grocery store wages and decent benefits. Anyone can do that, and only the overqualified will be passed up at an interview (barring major deficiency which is usually rooted out by screeners/recruiters).