Solved

is 90% of contracted ISP bandwidth really acceptable even when speed testing directly connected to their router?

Posted on 2016-08-08
27
99 Views
Last Modified: 2016-08-17
is 90% of contracted ISP bandwidth really acceptable even when speed testing directly connected to their router?
0
Comment
Question by:Xetroximyn
  • 11
  • 9
  • 3
  • +2
27 Comments
 
LVL 90

Assisted Solution

by:John Hurst
John Hurst earned 125 total points
ID: 41747436
If you are regularly getting 90%, that is acceptable.

You should be able to do point in time test directly at the router at 100%. If you cannot do this, then I would contact the ISP and ask them to explain.
0
 
LVL 28

Accepted Solution

by:
Jan Springer earned 250 total points
ID: 41747537
If I'm paying for X bandwidth and I consistently flat line at 90%, I call them on it.

You pay for 100%, you should have available 100%.
0
 
LVL 87

Assisted Solution

by:rindi
rindi earned 125 total points
ID: 41747558
The advertised bandwidth is a best effort and not guaranteed. There are many things that will reduce the bandwidth, for example how many others are using their services that are connected to the same distribution point. So if fewer customers use the internet at the same time in your neighborhood, the closer you'll get to the advertised bandwidth.
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747623
Your bandwidth depends upon your SLA (in general or specific) with your provider and nothing else.

Buying transit is different than buying broadband.

However, in best case scenario, you should be able to reach 100% even as a broadband customer.
0
 

Author Comment

by:Xetroximyn
ID: 41747653
Thanks everyone. To be clear this is a Viber internet circuit for a business. We supposedly bought a 100 Meg circuit but even when the carrier comes and hooks up a laptop directly to the router, the best thing yet is 92 Megs down and up.
To me that might be acceptable if there is other traffic on the line but this is with our entire network completely disconnected and only their laptop hooked up so there is no extra traffic.  They called this acceptable because it's within 90% and blamed it on TCP overhead. But don't speed tests take that into account? Like isn't the speed test actually telling me the number of bits that were flowing per second? Now of course when I'm trying to actually download a file there's going to be a parity bit and whatnot so I lose some of those bits to overhead too but doesn't the speed test actually tell you the total bits flowing over the circuit?
0
 

Author Comment

by:Xetroximyn
ID: 41747656
We are paying $2,000 a month for this circuit.....  I pay $70 at home and I get 220 mags down and 20 Megs up. Well that's what I'm contracted for but whenever I speed test I actually get 235 down and 23 up... seems like on a business circuit for a hundred Megs regardless of anything they are to be able to make my speed test say 100 megs
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747657
Do you mean "Fiber" instead of "Viber"?

If not, what type of circuit?
0
 

Author Comment

by:Xetroximyn
ID: 41747658
Err...  I said Viber...  Meant fiber...
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747659
I just read your comment.

Your provider needs to modify your configuration to handle any overhead and give you 100% to the premise.
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747660
Because, with your type of subscription you are not just a "broadband" customer but buying transit at a premium.
0
 

Author Comment

by:Xetroximyn
ID: 41747662
Another thing that struck me as odd is that they wanted me to set my router 2 100 full. When it was currently already Auto negotiating at 1000 full... I gave it a try and it made the speed tests even worse and so I refused to keep it that way and insisted that we set it to 1000 full for the negotiation.   From what they said though they apparently have it set to 100 full on the other side anyway. Which seems kind of ridiculous to me. Not that I have all that much experience with all this but it just seems crazy that we are paying $2,000 for a 100 Meg fiber circuit,  they have actual fiber coming into our building and they're using a 100 full negotiation at any point???
0
 
LVL 90

Expert Comment

by:John Hurst
ID: 41747664
Ask your ISP about performance guarantees. As noted above several places, your top bandwidth is not usually guaranteed. Ask them what you should expect.
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747666
Oh, bully.

They apparently don't know how to properly QoS your circuit and are forcing it at 100M by changing the speed.

Tell them, "no, hell no. fix your configuration."
0
Top 6 Sources for Identifying Threat Actor TTPs

Understanding your enemy is essential. These six sources will help you identify the most popular threat actor tactics, techniques, and procedures (TTPs).

 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747669
And, full disclosure, I set up and support regional service providers.  I am not speaking as an end user.
0
 
LVL 90

Expert Comment

by:John Hurst
ID: 41747676
Interesting. In 3 decades I have never gotten an average of 100% bandwidth for a sustained period.
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747682
It depends upon how your circuit is classified.

A lot of businesses will [essentially] purchase a residential class circuit.

And, it depends upon the provider.  But in a best case scenario, you should reach your subscribed bandwidth.
0
 

Author Comment

by:Xetroximyn
ID: 41747950
Thanks Jan!  My guess is that they will continue to blow me off.... any idea how I can really get them to abide?


(BTW - SLA Attached)
SLA-WINDSTREAM.pdf
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41747953
omg it's windstream.  let me read your doc.
0
 

Author Comment

by:Xetroximyn
ID: 41747991
I take it you have some history with windstream?

We have had a fiber circuit with them (20 meg) for 3 or 4 years.  Pretty stable... except once when we were completely down for over 24 hours.... Now we upgrade to 100 meg, and it's been a mess.... was supposed to be done in june and we still we didn't have our full upload bandwidth for the longest time until (they claim) today.  I still don't trust that we do though, because it says 80-90 sometimes and it says 20-40 sometimes.   And I know what's going out of my firewall is not enough bandwidth to knock a speed test that low.  During our peak production time (lots of SIP) maybe.... but not when we are not in peak production....
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41748052
Maybe  I read through the SLA too quickly but I don't see anywhere where it says that you buy 100% and we only have to give you 90%.
0
 

Author Comment

by:Xetroximyn
ID: 41748279
can you elaborate on this "Transit" vs "broadband" thing?  Is there a way I can be SURE we are a transit customer before I throw that in their face?
0
 
LVL 28

Expert Comment

by:Jan Springer
ID: 41748564
It's more transit vs business/residential.

Broadband customers are [copper] ethernet, ftth, dsl, or cable connected.

The difference is in the SLA and cost.

Based upon your cost and that you even have an SLA suggests that you are not a casual Internet customer.
0
 
LVL 16
ID: 41756219
Just thought I would log my 2 cents. An SLA is an SLA and the bandwidth guaranteed in the SLA is what you should be getting. There are certainly some technical differences in the various types of DIA circuits that might explain why you're getting less than the guaranteed, but it doesn't excuse the provider from providing what they guarantee. Fiber services usually employ policing/throttling policies. If you're getting less than the committed amount then the providers policing policies are not right. If you're doing something like Ethernet over Copper or Metro-E it could be a line capacity issue and the provider needs to bring in additional loops. Etc., etc., etc.

I will say that utilizing 3rd party speed test sites are not a reliable source, as those services can go through over-subscription periods that would skew your results. I generally always try to use the carriers only speed tests. Most of the good one's have their own.

MO
0
 

Author Comment

by:Xetroximyn
ID: 41757354
thanks!  FYI it's windstream.  It's a pure fiber circuit including last mile.  We have the fiber going straight to our server room.... capacity definitely is no issue here.  

I use speedtest.net (ookla).... windstream technically as a "speedtest.windstream.net" which is powered by ookla (says so right on the site).... funny thing?   Windstream's own test ALWAYS has the upload no higher than 20-30 megs while speedtest.net often has us up in the 80's.

When connected directly to their router speed tests reliably get to about 92.  With math about overhead the farthest they can get it down is about 95% (94.68).... then they blame the rest on browsers such that we ought to accept 90%+....

The thing is.... I can't seem to find documentation or a clear answer but it seems  completely absurd to me that one of the largest speed test providers (ookla) in the world would be so ridiculously incompetent to only count the payload data and not count all the header bits that they know are passing over the circuit in order for that payload data to pass across the circuit.  I mean am I wrong in that opinion?  Doesn't it seem like a speed test provider is being negligent if they send the 1460 byte packet and they know that 40 bytes are going and added to that to make it 1500 bytes and they don't count that as part of the calculation of how many bits went across the circuit in that particular second?  

They are in the business of telling you how many bits go across your Internet connection.   In my opinion they are not doing their job correctly if they are not counting the header bits in the calc as Windstream would suggest.  

Windstream's explanation of why speedtests cap out at 92megs even when done from a laptop that is only thing connected to their router.  
Thank you for bringing these points up.  I really appreciate this and want to address your concerns.  Windstream is most certainly responsible for providing you a 100 Mbps X 100Mbps circuit.  We have verified that we are delivering the subscribed bandwidth to your office.  We are testing expected speeds, both via Ookla and RFC2544 testing to the demarcation point.
 
The reason that we mentioned acceptable rate is that we are taking account for overhead in every packet as well as latency between your site and the speed test.  The 100 Mbps bandwidth provided is the maximum amount of data allowed over the link.  This does not account for TCP overhead, or for Latency on the span.
 
The standard IP MTU over the internet is 1500Bytes.  Out of those 1500Bytes, 20Bytes are utilized for IP headers and 20Bytes are utilized for TCP headers.  This leaves us with 1460Bytes for payload.
 
This equates to an efficiency of 97.33% (1460/1500).
 
If you add Ethernet (and VLAN tagging) into the calculation (see the calculations from Wikipedia here), then the throughput on a 100Mbps link is 100  * .9733 (TCP/IP efficiency) * .9728 (Ethernet efficiency) = 94.68 (Max TCP throughput on the link)
 
We also need to take into account latency to the speed test server, browser and flash versions  that can lead to inconsistencies while utilizing flash based speed tests which is why I offered 90% as an acceptable range.
0
 
LVL 16
ID: 41757809
When you're sitting behind your router/firewall and get the poor speeds have you checked your port speed on the outside interface of your firewall to make sure it's not connecting at half duplex? You router/firewall might be having issues autosensing and connecting at 100M Full Duplex or 1000M Full Duplex. You might have to hard code it if it's connected at half duplex.

MO
0
 

Author Comment

by:Xetroximyn
ID: 41759820
when my firewall is connected it connects at 1000full.

But to be clear - MY router was taken out of the equation.  The carrier disconnected our equipment hooked their own laptop up to their own router so that the laptop was the only device on their own router and the best speed tests they could muster was 92megs up/down.
0
 
LVL 16
ID: 41759847
If the carrier really locked you down to a committed rate of 100Mbps, realizing 92% of that is probably about right....plus or minus a percentage point. Different carriers do different things. I've seen some carriers exercise a little more lenience on the committed rate, so that the actual throughput is the contract amount, but that means that they police/throttle at a level a little higher than the contracted amount just so the end user can see that they're getting that "100Mbps" throughput. They probably do that, because they don't want to have to go through the task of explaining how the calculation of throughput works (e.g. MTU sizes, headers, etc.).

If you're getting 92% efficiency on your 100Mbps circuit then you're in good shape and likely in line with most of the public.

MO
0

Featured Post

Maximize Your Threat Intelligence Reporting

Reporting is one of the most important and least talked about aspects of a world-class threat intelligence program. Here’s how to do it right.

Join & Write a Comment

Suggested Solutions

Even if you have implemented a Mobile Device Management solution company wide, it is a good idea to make sure you are taking into account all of the major risks to your electronic protected health information (ePHI).
When it comes to security, there are always trade-offs between security and convenience/ease of administration. This article examines some of the main pros and cons of using key authentication vs password authentication for hosting an SFTP server.
Get a first impression of how PRTG looks and learn how it works.   This video is a short introduction to PRTG, as an initial overview or as a quick start for new PRTG users.
This video gives you a great overview about bandwidth monitoring with SNMP and WMI with our network monitoring solution PRTG Network Monitor (https://www.paessler.com/prtg). If you're looking for how to monitor bandwidth using netflow or packet s…

747 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question

Need Help in Real-Time?

Connect with top rated Experts

13 Experts available now in Live!

Get 1:1 Help Now