The most likely effect of having no QoS will be on what the party on the other end of the call hears, not what you hear on the Ooma end. You probably should try the upload test (listening to a loop-back echo call, or call in from another phone) in any event to see if you can live with the results you get with lots of upstream traffic. Oh well, that simplifies your life a little. I would definitely encourage you to "measure, make notes, tweak, rinse & repeat" - the advice from others (including me) might hint at what is possible, but they really will not have much bearing on the results you get because of equipment software differences and network loads. My experience will likely be completely different from yours however, as we have different routers with different QoS implementations & capabilities, different ISPs and geographic distances, etc. No wonder it worked - the upstream link wasn't saturated, it was only at 50%.Ĭlearly I still have a ways to go to get the behavior I want, and need to also measure downstream link saturation effects too. It turned my "No QoS, 70/5" link into a "QoS 40/2" link even without an Ooma call taking place. In my own case - using a particular priority scheme on a Cisco RV180 router, prioritizing the traffic from the Ooma only had no effect at all, even though QoS was turned on but (subsequently) associating all other traffic except for the Ooma to a lower priority class in fact did improve my call MOS scores and XC Latency under high upstream loads, but it did so in a way that "all other traffic" was clamped to lower throughput than what was available. (*2) QoS implementations on consumer routers are not uniform in their implementation or behavior. I'd recommend you actually measure your performance (make calls with heavy upload/download operations taking place) and proceed based on what you actually measure. And in many cases "I have no problems" is a paraphrase for "I never tested anything with any disciplined methodology, I just randomly do stuff and haven't really noticed a horrible loss of call quality". For some folks that may well be true, depending on their network equipment and radio environment, but it certainly isn't any kind of a guarantee for you as both of those will be different. (*1) You might well hear people say something like "I use it behind a wireless router with no problems". I'm on the west coast, and the round-trip time (my voice to the returned echo) appears to be at least one second. You can get an idea of how horrible the delays are by calling that test # I mention above and try the echo test. Transiting longer distances and more network hops are arguably only going to make latency and jitter worse as well. VOIP calls are universally worse than circuit-switched DS0 PSTN calls when it comes to latency. But this configuration caused other problems(*2), so beware: the devil is in the details. Once I got QoS working, my call MOS scores with an outgoing load returned back to about 4.3. This seems consistent with a saturated uplink. I could definitely hear drops in my (echoed) voice on the test call (option #3), whereas playing music (option #4) had no drops. The result? without QoS turned on, my MOS scores dropped from about ~4.3 to about ~3.3. One call was to a test call number (408) 647-4636, and the other to a friendly human. The purpose of the long-lived upload was to saturate the upload direction ( ~4.4 mbps in my case ) so that the Ooma upstream traffic would have to compete with other traffic on a saturated link. What I did was this: I uploaded a huge file to my Dropbox (an upload that would last for tens of minutes) and then made some calls. I just ran a test with my Telo sitting behind a Cisco RV180 router with QoS turned on. #2 You may want/have(*1) to experiment with QoS features on your router, if it has any. (There are other reasons, too: point-to-point Ethernet links are dedicated media, not a shared channel collisions & retransmissions don't occur in full-duplex wired segments, whereas they definitely occur in wireless). Simply put, the chances of problems are far greater with WiFi than a wired (Ethernet) channel. Not saying you are going to have problems if you link your router and Ooma via WiFi it's just that wired connections rely on technology which is (a) simpler, (b) faster, and (c) are almost completely immune from interference issues. RWCarr wrote:Should I do anything special to get best call quality? #1 if your integrated (modem+) router has any Ethernet ports, I would recommend a Ethernet connection to the Ooma box directly to your router (no intervening switches) each and every day of the week over a WiFi link.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |