QUOTE(andrew9292 @ May 9 2009, 11:06 AM)
I guess u had pretty much covered alot of aspects...but i have a few questions to add if u dont mind.
1.The Cogent CO engineer found out that the bottleneck is WITHIN TM. The thing is, is the bottleneck done on purpose or is it because of an overwhelming amount of users? All of sudden from Mid-March or so our internet started to crawl and they issued the news that they were having "circuit faults?" on SMW3 cable. Since then it appears that our connection never improved eventhough they claimed the circuit fault was fixed,
This leaves me to wonder if there really is a fault with the cable. And i also dont believe
all of a sudden in mid-March perhaps 1 Million users signed up for streamyx and
bottleneck its network until now? So what was the cause of the slowdown until now?
2.I have strong reasons to believe that TM has a technology to
control the routing/speed on a user-to-user or area-to-area basis, as after i complained to MCMC and PMO, it seems that my connection is getting back it's FULL speed on the SAME IP address. It was a huge change from that morning(slow) to evening(fast) on the same IP.
3. Best Effort and Fair Usage Policy need to be rewritten and MCMC has to put guideline on those 2 policies. Best Effort? It seems like they are not even putting any effort at all. Fair Usage Policy? Fair to who? All of us are getting dail up speeds when we pay nearly a hundred a month, some even pay more.
The event overview was the Cogent Co engineer confirmed the location of the bottleneck is due to the troubleshooting that i started to do first (we all know trying our best to solve the problems on our side is more effective). I couldn't trace from my side to our IDC (Internet Data Center), so i use another traceroute software finally pick up every single ip and the readings while having the IDC engineer trace from their ends and work with the Cogent Co engineers. The "uplink" and "download link" has route through different routing as eg, from PJ section 17 you go to Sunway using Federal Highway, and back to PJ using LDP via SS2. Usually, the "download link" will be having high traffic.
Now, we all know that TMNET has different customer base / group / classification namely "the normal user" who use dynamic IP and paying 66 / 88 bucks / month. And others who are fixed IP, IDSL and etc... In one of the email replied by the BMC department, let me show you the funny explanation that doesn't make sense and furthermore, the BMC manager was taking a shortcut by troubleshooting at his preferred location instead of come to my office. The funny things is, whenever we are facing slow browsing, i call their call center, they got no packet loss, connection is good. But on our side, the case is different and that time we were having 40-50% packet loss.
I've removed the identifiable information (in respect to others privacy). You have a look at the email, take a while guess, and you'll know what i mean.
Note: For those who have not gone through such process might not know what's happening in this event, you may skip this portion.Here's the emails.
===========
Good day.
Dear Sir and all..
Please be informed that we normally look into a very specific technical issue. Based on our understanding from the messages, the performance of the service is low. This kind of issue is catagorised as follows :
1) performance of the connectivity to all web / Domestic site
2) performance of the connectivity to International site
Based on the below email (before this) we have concentrated to 02 streamyx account given to us.
- xxx@streamyx
- xxx@streamyx
Basically the above are the source (source of the request packet) and I was make known that there is 01 international destination has been used for the test.
== destination : xxx.xxx.81.229
Both of the account have been tested from our site and there is no high latency or packet loss that may contribute to the said problem.
Anyway should we want to check about the connectivity to the Internatinal website , a destination can be used as benchmark to identify whether the existing performnace is acceptable or not , is there anything TM can do to improve the connectivity. For example, www.yahoo.com is used by many users to test the connectivity to US, www.google.co.uk for UK and so on.
Please also be informed that we have tried to come out with the possible return trace from the above IP given to us - xxx.xxx.81.229.
The result as follows :
Query: trace
Addr: 124.13.63.100
trace 124.13.63.100 (currently xxx@streamyx is connected)
Type escape sequence to abort.
Tracing the route to 124.13.63.100
1 gi10-0.224.core01.lax01.atlas.cogentco.com (66.250.4.5) 4 msec 0 msec 0 msec
2 te4-2.ccr02.lax01.atlas.cogentco.com (154.54.1.34) 4 msec 4 msec 0 msec
3 te7-3.ccr02.sjc01.atlas.cogentco.com (154.54.2.150) 16 msec 12 msec 12 msec
4 te3-4.mpd01.sjc03.atlas.cogentco.com (154.54.6.102) 12 msec 16 msec 12 msec
5 telecommalaysia.sjc03.atlas.cogentco.com (154.54.12.2) 328 msec 328 msec 324 msec
6 58.27.106.209 324 msec 324 msec 324 msec
7 58.27.103.122 312 msec 308 msec 308 msec
8 58.27.103.33 312 msec 308 msec 312 msec
The above is the best option for the return trace as of today 6.00.pm (malaysian Time).Should we see the result closely, the max latency during return trace to TM (domestic) site is averaging at 350ms which is almost equal to another US benchmark site such as www.yahoo.compls find the below : Pinging www.yahoo-ht3.akadns.net [209.131.36.158] with 32 bytes of data:Reply from 209.131.36.158: bytes=32 time=300ms TTL=55Reply from 209.131.36.158: bytes=32 time=359ms TTL=55Reply from 209.131.36.158: bytes=32 time=380ms TTL=55Reply from 209.131.36.158: bytes=32 time=403ms TTL=55Reply from 209.131.36.158: bytes=32 time=426ms TTL=55Reply from 209.131.36.158: bytes=32 time=449ms TTL=55Reply from 209.131.36.158: bytes=32 time=474ms TTL=55Reply from 209.131.36.158: bytes=32 time=291ms TTL=55Reply from 209.131.36.158: bytes=32 time=313ms TTL=55Reply from 209.131.36.158: bytes=32 time=336ms TTL=55Reply from 209.131.36.158: bytes=32 time=300ms TTL=55Reply from 209.131.36.158: bytes=32 time=382ms TTL=55Ping statistics for 209.131.36.158: Packets: Sent = 12, Received = 12, Lost = 0 (0% loss),Approximate round trip times in milli-seconds: Minimum = 291ms, Maximum = 474ms, Average = 367ms
The above result may differ from time to time because dynamic ip user will have different ip address every timenew PPP session is establish. That is why some users may have different experience. Anyway,
fixed IP address usersmay have better opportunity to gain better connectivity
.(depends on the prefered destination).Another issue raised was on email service used by customers. I would appreciate if you could furnish more info on thecomplainant. We will try to, at least isolate the issue and propose an alternative to minimize the impact tocustomer end.1) Streamyx login ID 2) email server name/IP/webserver/pop mail3) Technical support (from cust site)Thanks
Mohamad Faisal b Azizan
Broadband Management Centre,IPNOC
TM Wholesale
TM Bhd