Thread: 2am-3:30am/6am-8am congestion/high latency

Anyone else getting this ?

Wondering if it's a problem exclusive to BSNL or something that's happening to all ISPs.

The problem seems to be happening with airtel too.

It's also been happening for almost a month, with no status updates from BSNL.

Latencies during this time shoot up to 200-300ms above normal and then leading to connection choking=>packet loss.

As a result; unlimited night time plans are proving to be useless since downloading slows down to a crawl.

It's basically a bad or inefficient routing thing, or bsnl not giving the advertised bandwidth/contention ratio to users.

The following definitely aren't changing the problem:
1. changing dns servers - completely useless 5-10 ms difference
2. not a modem/splitter problem (changed both)
3. not a local line problem (had line checked)
4. traffic shaping utilities don't help.
5. restarting router doesnt change anything.

has a underwater cable broken ?
or is the bsnl plan changes affecting bandwidth during the night time period.

Me too used to get pure speed during night unlimited slot. I use UL 750. Haven't checked recently. When I last checked, my speeds were around 15 KBps.

Originally Posted by meetdilip

Me too used to get pure speed during night unlimited slot. I use UL 750. Haven't checked recently. When I last checked, my speeds were around 15 KBps.

Mine is fine!!!

Originally Posted by sangamam

Mine is fine!!!

I highly doubt anyone using bsnl isn't experiencing any congestion during 2-8am.

I'm yet to see pings/tracerts from people from around 2:30am and around 7:30am to international servers (EU/US/AUS/eastern asia) from anyone that claims their connection is fine.

In fact I'm pretty sure the problem exists with airtel too (going by their forums).

I was just looking for a confirmation and/or if there was an official sign or news article about a temporary problem like an undersea cable breakage.

@rsen
Kolkata has this problem
I am from Kolkata

not only from 2AM it's started from 9PM

Which IP are you in 117 or 59
I see the problem in 59

59 series.

Right now ping times to USA are fine (298-302ms avg)
ping times to australia are somewhat high (327 ms instead of normal 160-170)

I dunno if the problem is only in kolkata as the congestion/bad routing always seems to happen after packets leave India.

It seems like airtel also has the same problem; maybe they're using similar routing.

I tried to get Reliance BB but their sales team (the guys in charge of doing the groundwork) appears to be very sluggish. Although their central office and website say south kolkata is covered; over here they are "unsure".


here's a trace to a commonly used US server; cnet.com :

Code:

Tracing route to gtm-rb-tron-xw.cnet-basic-performance.akadns.net [216.239.122.40]
over a maximum of 30 hops:

  1    24 ms    15 ms    15 ms  59.93.192.1
  2    60 ms    60 ms    45 ms  218.248.255.26
  3   280 ms   280 ms   265 ms  115.113.128.17.static-mumbai.vsnl.net.in [115.113.128.17]
  4   279 ms   279 ms   279 ms  59.163.16.1.static.vsnl.net.in [59.163.16.1]
  5   279 ms   263 ms   280 ms  59.163.16.1.static.vsnl.net.in [59.163.16.1]
  6   420 ms   420 ms   421 ms  if-2-1.mcore5.NYY-NewYork.as6453.net [64.86.71.21]
  7   436 ms   420 ms   435 ms  Vlan1274.icore1.NYY-NewYork.as6453.net [209.58.62.17]
  8     *      468 ms   452 ms  Vlan456.icore1.NTO-NewYork.as6453.net [209.58.26
.41]
  9   296 ms   296 ms   293 ms  192.205.35.13
 10   296 ms   294 ms   296 ms  cr2.n54ny.ip.att.net [12.122.86.10]
 11   293 ms   327 ms   296 ms  cr2.wswdc.ip.att.net [12.122.3.38]
 12   294 ms   295 ms   295 ms  cr1.attga.ip.att.net [12.122.1.173]
 13   296 ms   296 ms   296 ms  cr2.dlstx.ip.att.net [12.122.28.174]
 14   294 ms   296 ms   294 ms  gar10.dlstx.ip.att.net [12.122.138.121]
 15     *        *        *     Request timed out.
 16   296 ms   295 ms   296 ms  c18-rb-gtm6-tron-xw-lb.cnet.com [216.239.122.40]
Trace complete.

ping times to this server are 298-301ish ms, which is normal; but as you said problems have already started at 9:25pm. Pings to the NYC mcoreicore servers are very high (i think tata owns these) and about 200-300ms above normal.

Between 2-8am you will see a traceroute similar to this but in additional all the pings to hops after nyc i.e. AT&T servers are all 450-500+; leading to an overall ping to any US server being 200-300ms above normal. There is also connection choking, low bandwidth=>slow speeds when this happens.

What sucks is vpn or any other kind of service becomes impossible lol.

@rsen
it's a local fault mean fault in Kolkata , I dont see 117 this problem.
When you face this problem ( 1st August was very bad ) even you cant open the usages records address, that pages also take time to load or even some times cant loaded.
I also check the bandwidth, the download speed on P2P is very good, but only effect on http, if there is cable problem or fault on cable ( underwater cable broken ) then all protocols will not working good (P2P)

If some one starts ddos attack or http protocol attack from his/her home to the main server of kolkata then this problem can be seen.

I also check on BSNL IP 61, which is giving by dialup service, working good, no packet loss and no any kind of problem in http.

Same problem for me too.I am using the 59 series in Kolkata but i have applied for the Rel WiMax.I had enough of this BSNL crap.Paying almost 900/- ( UL 750 ) each month from 3 years but not more.

@rsen

I have been reading some of your posts for quite some time and honestly you are the only one who has raised some grave concerns regarding the ongoing latency problem which after talking to some people using different ISP's I found to be true. I think Indranil is correct and that the problem is with users for the 59 series bcoz I am from the 117 series and have not experienced this problem yet. However, I should mention that I didn't check with any tracer or ping test softwares about the problem,but, which I am going to do soon. For a general idea, I think you mentioned that this has been the case since July (almost the whole month) and now in August at the time aforementioned. I download a lot of stuff from different servers based in EU and sometimes in the US. To give an idea, below is the screenshot of my downloads in July and August (13th) and I should mention again that ALL these downloads have been done during the NU period.

This is a very genuine concern raised by you and not only entails BSNL users but also other ISP providers. After significant searching I did not find any news of cable repairs anywhere so I guess this has got something to do with ISP's in India. Will follow-through with ping tests and tracers shortly.