Results 1 to 3 of 3

Thread: QoS questions

  1. #1
    Join Date
    Feb 2018

    Default QoS questions

    Brand new at trying to understand this. We're looking at Bandwidth Monitor to throttle audio and video streaming in out environment. Per the documentation and videos, QoS needs to be enabled to use Bandwidth Monitor.

    We already have Bypass Rules in place in Config > Network. We want to apply Bandwidth Monitor to a single rack (for now - default rack later).

    From reading, QoS will apply to the traffic that's bypassed. Those rules are set in QoS Rules and should match, one-to-one, the rules already set in the Bypass Rules. What happens, though, when you don't have any QoS rules set? Does that traffic automatically get rated above everything else?

    Then, on the linux stack side, using Bandwidth Monitor for the non-bypassed traffic, when you enable Bandwidth Monitor on one rack but not another, what happens to the traffic on the non-enabled rack?

    For example: I have an untangle setup that bypasses allows traffic into a web server. I have 2 racks - A and B. Looking at the configuration in documentation and on the demo site, I want to prioritize all HTTP and HTTPS traffic as High (or Very High) on rack B.

    If I:

    enable QoS but don't set any QoS rules on my Bypass Rules
    In rack b, apply Low priority to audio and video streaming
    In rack b, apply High (or Very High) priority to HTTP and HTTPS

    what happens to the:

    bypassed traffic that doesn't have any rules and the traffic in rack A?

    Does Untangle prioritize the Rack B traffic above the traffic in Rack A? Does Untangle prioritize the Rack B traffic above the bypassed traffic?

    I'm wanting to try to understand this before I apply it to a live environment. Thanks so much for any insight that anyone has.

  2. #2
    Untangle Ninja jcoehoorn's Avatar
    Join Date
    Mar 2010
    York, NE


    This won't end well.

    QoS isn't about total packets or throughput any more. It worked that way somewhat when you had to download everything in full, but in the era of streaming media it's not that effective. Instead, QoS is about guaranteeing service for packets that are time senstive. That means streaming services need a higher priority when you use QoS.

    The best you can hope to do is leave the priorities at the default, but create reservations for business-important traffic, so that streaming services can't crowd it out. But that's about identifying what is really important, rather than limiting what isn't. This will be especially important once you realize how many business-related services use common streaming protocols: WebEx, VoIP phones, user training products...

    If you didn't have enough bandwidth before QoS, you're unlikely to have enough after.

    As for the specific questions:

    Once you enable QoS, all traffic is prioritized. Traffic that doesn't match any rule is left at the default (medium) priority.

    So if you "enable QoS but don't set any QoS rules on my Bypass Rules", the traffic all has the same (medium) priority.

    In rack b, if you apply Low priority to audio and video streaming, that traffic will have Low priority, and will therefore probably become functionally broken any time you have even mild congestion, with constant buffering. You just can't de-prioritize time-sensitive traffic and expect it to work.

    In rack b, if you apply High (or Very High) priority to HTTP and HTTPS, that traffic will have a higher priority... but you won't notice any difference. Priority only matters when there are queued packets, and HTTP/HTTPS traffic is not very time sensitive. It won't matter to the user's perception of loading a web page if the http request or response is delayed a few extra milliseconds. Additionally, you just shouldn't do this. About the only traffic that should ever be prioritized as "Very High" is DNS.

    Finally, in both cases, rack A traffic would remain prioritized at the default (medium) level.

    So what good is QoS, then?

    Well, assuming you do have enough bandwidth (even if just barely), QoS can make sure your time-sensitive traffic (likes streaming services) continues to function well in cases when you have bursts of moderate to severe congestion.

    It can also be used to de-prioritize heavy traffic that is not time sensitive. It won't matter if a large download that's gonna take a few minutes anyway takes just a little longer. That can help improve your bandwidth use, which is why many look at QoS as a bandwidth management tool. Unfortunately, this type of traffic is much less common than it used to be. More and more services use streaming media vs downloads, and so QoS has much less impact for actually managing total bandwidth.

    Finally, QoS can be used as a tool to control unwanted traffic. For example, bittorrent traffic can be next to impossible to eliminate completely in some situations, but Application Control+Bandwidth Monitor+QoS can throttle it down to an insignificant nuisance, where the protocol still believes it works and so doesn't bombard your network, but end users aren't able to accomplish anything significant.
    Last edited by jcoehoorn; 02-13-2018 at 01:40 PM.
    Five time Microsoft ASP.Net MVP managing a Lenovo RD330 / E5-2420 / 16GB with Untangle 14.1.1 to protect 500Mbits for ~400 residential college students and associated staff and faculty

  3. #3
    Master Untangler abailey's Avatar
    Join Date
    Mar 2016


    This is the way I understand it:
    Under QOS you will see a Default Priority, That priority is given to all traffic that does not match a QOS Rule or a Bandwidth Monitor rule. QOS rules only effect bypassed traffic and Bandwidth Monitor rules effect non-bypassed traffic that matches your rules in the rack.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

SEO by vBSEO 3.6.0 PL2