dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
1115

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

Anyone else "downgraded" to 3 channels?

Just noticed I now have 3 bonded downstream channels. Used to have 5. Wonder if they rolled back a change here, or if they are tweaking the configuration and only bonding 3 channels for regular tier.

Speeds are still fine.

Tradewind
Premium Member
join:2005-11-08
Marlboro, NJ

Tradewind

Premium Member

That'll teach you for sticking up for CV
frdrizzt
join:2008-05-03
Ronkonkoma, NY

frdrizzt to jaa

Member

to jaa
Did you get a new IP? If so, they may have had to move your node temporarily to an older cmts that does not support 5-channel bonding.

There could have also been an issue on the cmts that caused them to shut off the additional frequencies temporarily. But they wouldn't just take the ability to bond those channels away for no reason when there is such a push to move nodes over to that very setup.

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

Yes, my IP did change.

I thought maybe they changed "regular" tier to 3 channels only, to ensure higher-tier customers get some reserved bandwidth - which I think makes sense. At least until they get more downstream frequencies - I think they should have at least 12.

EliteData
EliteData
Premium Member
join:2003-07-06
Philippines

EliteData

Premium Member

with analog tv gone, there should be plenty of frequencies available for both tv and internet.

bohratom
My Jersey Giants finally winning again..
join:2011-07-07
Red Bank NJ

bohratom to jaa

Member

to jaa
Long as your speeds stay the same I would'nt fret about it. 5 channels can still be congested if they oversubscribe to.

blohner
join:2002-06-26
Lehigh Acres, FL

blohner to jaa

Member

to jaa
I like the 12 channel Idea (still 5 here today with rock solid 60/8 on boost+) . I think what is equally important - and depending on topology not always simple - is making passed houses (potential customers) per node smaller and smaller... I had one 3 month lasting peak slowdown in the 10+ years (with a 1 year FIOS break) I am with OOL - in the same time my node has been split times (maybe even more - what I can see is the node that connections to 4 cross roads used to be one, then two, now four)... Adding that to going from 1 or 2 channels originally to 5 now makes it a total of 20 channels for the same number of customers.... As the plant becomes able to handle higher frequency and with the demise of analog it will be easier to add more and more channels and stay on par with all but the most extreme FIOS offerings...
cablewizzard
join:2009-06-14
Woodbury, NY

cablewizzard to jaa

Member

to jaa
said by jaa:

Yes, my IP did change.

I thought maybe they changed "regular" tier to 3 channels only, to ensure higher-tier customers get some reserved bandwidth - which I think makes sense. At least until they get more downstream frequencies - I think they should have at least 12.

So how's your speed?

We should widen the LIE to 12 lanes, too, or build a second route altogether, but I doubt taxpayers will be willing to pay for it, cause: we don't do strategic infrastructure builds even 5 years ahead of actual need in the U.S. of A. We let things run to shit first, and then complain loudly about our own failures and failing leadership.

16-channel DOCSIS equipment is not viable in the market space today, it's bleeding edge with no current "general availability" product in the market, especially no VOIP-capable one. Broadcom (at CES) demonstrated some 16-channel equipment - stitched together with TWO 8x4 chips: see what I mean?

As both your IP and your available DS channels changed, your node might have been moved to a different CMTS, to balance out traffic and system load. The CMTS may be only built out for 3 DS channels: headends can have many CMTSs, and it's unlikely that all are upgraded at the same time.

Again, all this is highly speculative - the additional channels may be down over a network or RF issue (a single impaired, high-error-rate channel in your bonding group will slow down actual speed significantly). Might have split bonding groups for traffic management, too, but that would be specific for 4x4 vs. 8x4 modems.

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

If you read my original post, you would see that speeds are fine.

12 channels would work fine with the existing equipment.

2 shared with regular and boost.
2 shared with regular and ultra.
4 shared with boost and ultra.
2 dedicated to boost.
2 dedicated to ultra.

4 bonded for regular, 8 bonded for boost and ultra.

3 channels for upstream - one per tier.

Your LIE analogy is way off. There is no need for "widening" - the HFC network stays as is. No new nodes, no new fiber, no new coax. It is merely a re-allocation of frequencies, and perhaps some equipment upgrades.
cablewizzard
join:2009-06-14
Woodbury, NY

cablewizzard

Member

said by jaa:

If you read my original post, you would see that speeds are fine.

12 channels would work fine with the existing equipment.

2 shared with regular and boost.
2 shared with regular and ultra.
4 shared with boost and ultra.
2 dedicated to boost.
2 dedicated to ultra.

4 bonded for regular, 8 bonded for boost and ultra.

3 channels for upstream - one per tier.

Your LIE analogy is way off. There is no need for "widening" - the HFC network stays as is. No new nodes, no new fiber, no new coax. It is merely a re-allocation of frequencies, and perhaps some equipment upgrades.

Going to 12 DS channels is widening on the CMTS side, and those don't come free : they eat up capacity for VOD (which is quite critical), general video-QAMs, and special services like Optimum Homes/Cars/News interactive channels.

I've broken this down in a thread here last year, but joining all available DS channels across all tiers is better economically, and will show better total system utilization for all users: general Erlang calculations apply here:

I don't think there's enough Boost/Boost+/Ultra users to justify dedicated channels anymore (especially with the small number of US channels) - they just push system cost incurred for the non-standard users way above acceptable levels, relative to the additional revenue CV makes of such premium customers (which as you know are very modest: BOOST+ = 300% more bandwidth for only a 30% price premium over OOL) .

With Erlang, this works both ways - not only do regular OOL customers get to use all system capacity if required, but a small number of Ultra customers (like: 2 running at full bore) clogging ALL channels available to them will benefit if other channels have (at least some) capacity.

I wouldn't QoS premium users over regular OOL users either, even if all channels were shared: this immediately gives an unbeatable advantage of one group over another - an advantage where just one handful of highly-consuming users (from either side of the fence) can make the other side very unhappy very quickly.

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

Therein lies your problem. You are thinking like an engineer, talking about things like total system utilization. Let the "up to" nonsense apply to the masses on the regular tier, and give awesome performance to the premium (highly profitable) tiers.

Boost, boost+, and ultra are extremely profitable. While the price increase is modest, it is all profit. And gets the high-bandwidth users off the regular tier.

I'm sure CV could tweak the allocations based on actual usage by tier - I have not idea what it is. But premium tiers should get priority.
frdrizzt
join:2008-05-03
Ronkonkoma, NY

frdrizzt

Member

I'm sure CV would rather have 6 OOL users than 1 ultra user, as the ultra user has a lot bigger impact (maxing out his connection would be the equivalent of the other 6 using their full connection at once, roughly, which is much less likely to happen). Plus, an Ultra user only generates twice the money as an OOL customer, and uses more bandwidth, necessitating more frequent upgrades on the node (more channels or additional node splits) to maintain with the ever-growing bandwidth usage. Ultra is best for advertising & for keeping customers (mainly businesses or home-offices) who would change providers for competitors who offer high speeds, which could cause them to lose a number of business phone lines at $30 a piece (if over 4), or residential TV package & phone line if they bundled with a competitor. I can't imagine they make a killing off the average ultra user (who is obviously not just an average bandwidth utilizer), but instead that it would cost them more to not have that as an option. In a nutshell, I think there is more to lose in focusing primarily on the much smaller number of Ultra customers than the standard OOL.

This post contains opinion, and may be totally out of line with the reality of what CV execs are thinking/planning.

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

I'm sure they would rather have 5 OOL users and 1 ultra user than 6 OOL users.

Tier does not necessarily indicate how much bandwidth is used, and bandwidth is a small part of CVs cost structure.

The point is they need to allocate more frequencies to internet - a lot more. They can fight all they want, but eventually they will just be a pipe - my box will connect directly to the thousands of content providers through the internet, just like it did OTA before CATV.

aannoonn
@optonline.net

aannoonn

Anon

said by jaa:

Tier does not necessarily indicate how much bandwidth is used

So then why are you saying CV should allocate bandwidth by tiers?

jaa
Premium Member
join:2000-06-13

jaa

Premium Member

said by aannoonn :

said by jaa:

Tier does not necessarily indicate how much bandwidth is used

So then why are you saying CV should allocate bandwidth by tiers?

Apples and oranges. The tier a subscriber is on does not necessarily indicate how much bandwidth they use. Allocation of bandwidth across tiers (12 channels with max 8 bonded means you have to do an allocation) would be based on usage by tier.
TheWiseGuy
Dog And Butterfly
MVM
join:2002-07-04
East Stroudsburg, PA

TheWiseGuy to cablewizzard

MVM

to cablewizzard
said by cablewizzard:

I've broken this down in a thread here last year, but joining all available DS channels across all tiers is better economically, and will show better total system utilization for all users: general Erlang calculations apply here:

Did you take into account that D2 users will be moving between channels via DCC in that analysis. I understood that if you were taking only D3 Modems into account that using all channels for every user would give better utilization but how does a heavy D2 user fit in, doesn't he increase utilization only on one channel? Can D3 users be Dynamically assigned channels and is it possible that in order to load balance heavy D2 users, reducing the number of bonded channels except when utilization is high and moving heavy D2 users off the bonded channels makes sense?