extremes connected by ethernet 160mbps, 24mbps with WiFi..why?

Discussion in 'Mac OS X Server, Xserve, and Networking' started by gavcav, Feb 1, 2017.

  1. gavcav macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #1
    i've spent all day trying to switch my airport home network from being wirelessly connected to connecting by ethernet to create a roaming network, but it's not worked as i expected.

    i have 2 of the latest extremes and 2 express's..i wanted to connect 1 extreme to my modem via ethernet in the lounge and then connected to another extreme via ethernet in the hallway. from here i wanted 2 express's in the kitchen and 1 bedroom connecting wirelessly.

    firstly it was a real pain connecting to my modem, then secondly when i checked the download speeds i was really shocked..connected via ethernet i was averaging 30-35 Mbps, and with Wifi i was getting 160 Mbps!..seeing as i'm hoping for a wired network this was really disappointing and more..totally not expected..why is the wired connection giving me a slower download rate?..can anyone help me out please?
     
  2. BrianBaughn macrumors 601

    BrianBaughn

    Joined:
    Feb 13, 2011
    Location:
    Baltimore, Maryland
    #2
    If your modem also a router? If so, is it in bridge mode? What is the configuration of each Extreme?

    Side question...the Expresses in kitchen and bedroom. Are these rooms progressively further away from the Extremes? I can say that if your Express can get a signal in the bedroom then it should be to extend the signal beyond the bedroom as any device in the bedroom can most likely get the same effective wifi speed as the Express.
     
  3. belvdr macrumors 603

    Joined:
    Aug 15, 2005
    #3
    Start with the modem, per the post above. Then start with only the first Extreme and test. Add the second Extreme and test again.

    Does the Ethernet speed change when connected via wired in either Extreme?
     
  4. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #4
    1st, Roaming networks and Extended Networks are different beasts. From your description, you are trying to do both.

    Roaming networks use Ethernet to connect all of the routers and Access Points (AP) to the LAN, with the Network > Router Mode DHCP & NAT for the router only, and Off (Bridged Mode) for the remaining APs. Then, on the Wireless > Network Mode, all should be set to Create a Wireless Network with the network name and password the same for all APs and the router. If using dual mode (2.4GHz and 5GHz), if you name the 5GHz network something different (NETWORK and NETWORK-5G for example), you can choose which to join based on which gives better performance in a given area (5GHz has shorter range but faster speeds generally).

    If the second Extreme, connected by Ethernet, is not set to Bridged Mode, you are creating an infinite loop between 2 routers and all kinds of ugly things will happen, including possibly slowing down the Ethernet speed for the connected devices on either of the Extremes as the routers fight to figure out how to route the traffic. More likely, your routers will constantly reset to clear the routing cache which is learned dynamically using discovery packets.

    If the wirelessly connected Expresses are set to Create A Network, they will have no place to route the traffic and you will have a closed loop for devices connected to these APs. If you want these to connect wirelessly, choose Wireless > Network Mode Extend A Wireless Network. However, understand that Wirelessly extending your network will diminish bandwidth by roughly 50% for devices connecting to these extended APs as the WiFi is split between client access and backbone access to the router. If you go this route, be careful to place these far enough from each other so that the router's signal is clearly the strongest so they will prefer routing packets to the router and not through each other which will further reduce speed on the Expresses. If it is at all possible, either don't use these extra AP, or if you cannot get WiFi signals to those areas, use Ethernet, or Powerline Ethernet, or MOCA to extend the wired LAN to these locations and then follow the Roaming method above.

    Keep in mind, the goal is to use a minimal number of WiFi access points to limit the amount of interference in your space. WiFi is a rather promiscuous network access technology, sending wireless packets to every host (AP or client) in range (even if not on the same SSID), then dropping anything it is not supposed to handle. The more AP you have, the more interference and thus slower speeds. If the distance between APs is great enough to avoid too much overlap (only possible with the Ethernet connected Roaming method), the interference will be minimized and speeds will be better. But, in Roaming networks (all AP connected via hardwire), the backbone speed on the wired LAN more than makes up for the interference and performance will be adequate even if there is overlap.
     
  5. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #5
    thanks brian, yep the modem is part of a modem/router hub which is switched to modem mode only. and i think i get your point regs the express's..they have to be a suitable distance apart as not to be swallowed up by the other airports?
    --- Post Merged, Feb 4, 2017 ---
    thanks..will do, i've not had time as yet as i'm working all weekend so i can't check regarding your point about checking the speed relative to each extreme.
    --- Post Merged, Feb 4, 2017 ---
    brilliant write-up techwarrior, you've made some good points i hadn't thought about. i have planned to ethernet the 2 express's at some point in the future, but i may keep them out of the loop for now and just set up the 2 extremes which thinking about it, should give wifi coverage around my house.
    i have noticed that devices were clients to more than one airport so i guess this is not ideal?

    on another front..what do you think about powerline/homeplug ethernet connections to the extremes?..is it preferable to have a single dedicated ethernet cable between the 2? or does the use of powerline/homeplugs do just as good a job? also, if i decided to introduce a gigabit ethernet switch at some point in the future, where is the best place to introduce it..after the primary extreme or before the primary extreme?
    --- Post Merged, Feb 4, 2017 ---
    update: my network is at the moment an extended wireless network ..checked this evening and i'm getting quite low download rates now..around 40mbps at the computer desk in the hall and in the lounge..my ISP should be giving up to 200mbps and before i introduced the extremes, using my ISP hub i was averaging around 160-170mbps.
    i've got the primary router at my desk, also connected to my imac by ethernet and i'm achieving the 160-170 mbps..with my using the extreme connecting by ethernet to my computer reduce the wifi perfromance?
     
  6. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #6
    You may be better off without the expresses until you can connect via Ethernet. I don't see an issue daisy chaining the Ethernet, the wired connection is a network switch in the Express, so it should work just fine. As for power line, I have good success with the TPLink TL-PA8010P KIT which claims 1200Mbps. I think performance will vary depending on how many circuits you cross and distance, but clearly this unit exceeds my ISP speed (50Mbps).

    My config is similar to yours, a current generation Time Capsule for the main router <> GB Ethernet switch <> an n-1 gen Extreme on the opposite end of the lower floor connected via GB Ethernet to the GB Switch. I also have a dual mode Express upstairs connected over power line to the GB switch, then finally an older Express with WiFi disabled connected by Ethernet directly to the GB switch (this is only used for Airplay to power a speaker via Aux to the outside patio).

    As for a GB Ethernet Switch, hard to say where it would be best placed in your network, but clearly after the Router. Clearly, the switch should be located where Ethernet cable runs make sense.

    As a general rule, the number of access points should be the minimum needed to adequately cover your space. The 2.4GHz networks are particularly susceptible to contention, I went out of my way to look at the channels used by my 3 Access Points, as well as neighbors and hard set my 2.4G channels to avoid conflict. This had very positive results on the 2.4G speeds, but as 5G has many more channels, it was not necessary to do on the 5G frequencies. I am able to get 50-60Mbps on 5G everywhere in my home, and 20-55 Mbps on 2.4G. I name my 5G SSID different than my 2.4 so that devices that can take advantage of the 5GHz can choose to connect to the faster network.
     
  7. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #7
    Thanks, I'll have a mess around with my network tomorrow...out of interest, how would I view what devices are on which channels of my network and how to hard set them?..regs gavin
     
  8. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #8
    ..i don't fully understand when you mention "daisy chaining the ethernet, the wired connection is a network switch in the express"?
     
  9. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #9
    ok..THIS IS BIZARRE!..i've no idea what on earth is going on here?

    STARTING SETUP - connection by wifi
    with both extremes connecting by wifi (extending?)...i tested the wifi between the 2 extremes and i was getting at my work desk in the hallway 39.96 mbps download/10.47 upload....in the lounge i was getting 41.86 mbps download/6.74 upload..Ok but still disappointing since my ISP provides 200 mbps.

    ethernet/homeplugs(powerline)
    after setting the lounge extreme to factory settings i connected up both extremes with an ethernet connection via homeplugs/powerline adapters i tested again at my work desk in the hallway and got these readings..29.80 mbps download/10.44 upload...in the lounge i got 1.24 mbps download/1.03 upload!!!

    imac connected to primary extreme by ethernet
    my imac is connected to the primary (router) extreme by ethernet and i'm getting 164.79 mbps download/10.30 upload which i expect with my ISP giving me up to 200 mbps

    cat63 cable connected between each extreme
    so i threw a cat6e cable between them both and got 46.74 mbps download/10.46 upload in the hall and 48.21 mbps download/10.26 upload..so roughly the same as i was getting via wifi before i started!..am i missing something here? i thought the ethernet connection would be much better than wifi but i seems there is hardly any benefit at all?

    i'm totally lost here so if anyone can help i'd be so appreciative!..regs gavin
     
  10. BrianBaughn macrumors 601

    BrianBaughn

    Joined:
    Feb 13, 2011
    Location:
    Baltimore, Maryland
    #10
    For me there's too much scattered data to make any sense of this and I'd need a diagram with locations and precise info on the test connections to even begin to analyze. No need to respond to me...I'm just sayin'.
     
  11. belvdr macrumors 603

    Joined:
    Aug 15, 2005
    #11
    Forget the WAN at this point. Test speeds internally to see if there's an internal network problem. I use iperf to do this. It requires an install on two separate systems (a server and client).
     
  12. gavcav, Feb 6, 2017
    Last edited: Feb 6, 2017

    gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #12
    Screen Shot 2017-02-06 at 14.26.10.png Screen Shot 2017-02-06 at 14.17.07.png
    ok brian i will reply..to simplify, in my hallway i have a desk with my imac, ISP hub (in modem mode) and the primary extreme connected to it by ethernet.i also have a netgear plp1000 powerline adapter connecting to my primary extreme.PINK CIRCLE SHOWS LOCATION

    in the lounge, i simply have a secondary extreme with another netgear powerline plp1000 powerline adapter connecting it by ethernet.GREEN CIRCLE SHOWS LOCATION

    the distance between the 2 extremes is approximately 25m tops..please see attached image of layout.. Screen Shot 2017-02-06 at 14.17.07.png
    --- Post Merged, Feb 6, 2017 ---
    thanks..how do i do this?..sorry for the ignorance but i'm simply your standard house owner doing a DIY install
    --- Post Merged, Feb 6, 2017 ---
    ok..so i'm getting these speeds with these connections..

    STARTING SETUP - connection by wifi

    hallway 39.96 mbps download/10.47 upload
    the lounge 41.86 mbps download/6.74 upload

    ethernet/homeplugs(powerline)
    hallway 29.80 mbps download/10.44 upload
    the lounge 1.24 mbps download/1.03 upload

    imac connected to primary extreme by ethernet
    164.79 mbps download/10.30 upload

    cat63 cable connected between each extreme
    hallway 46.74 mbps download/10.46 upload
    lounge 48.21 mbps download/10.26 upload

    to summarise i'm getting basically the same speeds with wifi and ethernet which doesn't make sense to me having up to 200 mbps from my ISP?
     
  13. belvdr, Feb 6, 2017
    Last edited: Feb 6, 2017

    belvdr macrumors 603

    Joined:
    Aug 15, 2005
    #13
    It appears to me, based on the drawings only and the Pythagorean Theorem, that the two Extremes are more like 5-8 meters apart. This is way too close, unless each room is a Faraday cage.

    Download iPerf here:

    https://iperf.fr/iperf-download.php#macos

    Make sure to use the same version on all systems.

    Here's a use guide:

    https://www.macobserver.com/tmo/answers/how-to-test-speed-home-network-iperf/page2

    Post the results for both wired and wireless here. I'd run the server only on a wired connection.

    This will tell us if you have an issue internally or externally. If this is too complicated, I'd suggest the trial-and-error approach and start with only one Airport Extreme and see how things change.
     
  14. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #14
    thank you very much...i'll try and follow the instructions and get back with the results..regs gavin
     
  15. belvdr macrumors 603

    Joined:
    Aug 15, 2005
    #15
    See my edit above too. Looking forward to the results. :)
     
  16. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #16
    I took your question above whether you could run Ethernet to one Express from the router or GB switch, then use the LAN to WAN ports on the first Express to connect Ethernet to a second Express. This should work just fine, the WAN\LAN ports on Express are effectively an Ethernet switch, and when APExpress is not the router (i.e. in Bridged mode), the WAN port basically becomes a LAN port.
     
  17. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #17
    This weekend, I took a closer look at my Airport setup and found some interesting results.

    My setup: Time Capsule (ac) as main router in the basement office, connected by Cat5e to cable modem with Comcast 50Mbps service. On the far end of the basement, AP Extreme (dual-band n) connected by GB Ethernet to the Time Capsule. AP Express (dual-band n) upstairs connected by TPLink 1200Mbps Power Line adapter to the Time Capsule.

    All Airports were set to dual mode (2.4 + 5GHz), setup as Roaming (Create a Wireless Network), and all were set to automatic channel selection on both 2.4 and 5GHz bands. In reality, all three were using channel 6 for 2.4GHz. Two neighbors have Wi-Fi networks broadcasting 2.4GHz signals in the channel 1-9 range but were weaker signals than my own network. The neighbor's 5G signals were virtually non-existant. With this setup, I got internet speeds of 50-60Mbps on 5GHz everywhere, and 8-60Mbps on the 2.4GHz band.

    Next, I staggered the 2.4GHz channels on the three Airport devices so one was set on Channel 1, another on Channel 6, and another on Channel 11. I then re-tested both bands, 5GHz remained unchanged and 2.4GHz ranged from 24-60Mbps where the lowest speeds were in locations where my neighbor's signal was the strongest.

    My take, the overlapping channels on my own Airports were causing collisions\interference with each other and my neighbors signals weren't helping matters. As I understand, each channel range is a separate collision domain. Hosts in a given collision domain must wait in queue for access to use the channel. The more hosts in a given channel range, the contention for access to use the channels increases, thus speed slows. Since packets are sent to all radios in a given channel range, even foreign networks, the contention increases when neighboring networks and hosts operate in the same range of channels.

    To analyze where your bottlenecks reside, start with the physical LAN by connecting a Mac\PC to an Ethernet port at each Access Point location and test the speeds using something like speedtest.net and\or LAN Speed Test (see below). The results should be similar in all locations, else you have a weak Ethernet link between that location and the router. Ideally, LAN speeds should be equal or greater than your ISP service.

    Next, test 2.4 and 5GHz Wi-Fi speeds The easiest way to do this is to use a different SSID (network name) on the 2.4 and 5G bands (ex: NETWORK and NETWORK5G) on all of your Airport devices. Then, test and note results on each frequency while connected and close to each access point. Results should be relatively consistent on 5G given interference should be minimal. For the 2.4GHz band, expect varying results.

    Then, using AP Utility, spread the channels on all of the Airports so they avoid overlapping. Repeat the Wi-Fi testing. Finally, if neighbor networks appear in your home, figure out which 2.4G channels they are using and set the Airport channels so that the overlap occurs on the Airport device where the neighbor's signals are weakest.

    The tools I used were:

    LAN Speed Test Lite - free tool for Mac or PC, for testing local network connection speeds available here. Note that this tests file transfer speeds, disk read\write speeds impact the test results. To get network speeds accurately, there is a combo of server and client you can purchase for $11 which tests without reading\writing from disk.
    Airport Utility - to verify my Mac or iPhone was on the list of connected devices on the Airport device I was testing.
    Speedtest by Ookla - free in the ATV and iOS app stores, and Speedtest.net on PC or Mac via browser to test speeds throughout the home to internet locations.
    Wi-Fi Scanner ($19.99 in the MAS here). This is a good tool for visual representation of the Channels used by all networks within range. If you have a portable Mac or PC, all the better as you can test in all locations. Any Mac or PC WiFi analysis tool will work, the key is being able to see which channels are in use at each location to find the overlap.

    Final thoughts: avoid using 2.4 GHz if possible, it is most prone to contention. If you must use 2.4GHz, limit the number of Access Points and hosts you use on this frequency, and avoid channel overlapping. And, if your only option for connecting additional Access Points is wirelessly to the router, only use 1 device connecting this way or your entire 2.4GHz band will suffer. Some experts claim more than 1 wireless extender will absolutely clobber your entire Wi-Fi network.
     
  18. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #18
    Thank you for all that!...so basically, the channels were too close and causing interruptions?..where in utility can I view/set channels? I must say I've not noticed that option there?
    Regs Gavin
     
  19. techwarrior, Feb 6, 2017
    Last edited: Feb 7, 2017

    techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #19
    Not in Utility (for viewing that is). You can set the Wi-Fi channels on Wireless > Wireless Options.

    You will need some sort of analyzer such as the one noted at the end of my last post. There may be freebies out there? Key is to collect info on not only yours, but neighbors channels too. And, use 5GHz if at all possible.
     
  20. belvdr macrumors 603

    Joined:
    Aug 15, 2005
    #20
    That's why I disabled 2.4 altogether on my network. My devices support it and I got tired of fiddling. I no longer have any wireless issues.
     
  21. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #21
    Unfortunately, older devices like IP Cameras rely on 2.4GHz. I didn't realize, but with Sierra's AP Utility, there is now an option to disable 2.4GHz. On the channel selection list there is now an "Off" option. I might have to try that, disabling 2.4 on all but one of my AP and let that one AP operate in a channel range completely outside of my neighbors.

    With different SSID, you can choose to ignore the 2.4GHz network and always use the 5GHz on devices that support it, and no negative impact by having 2.4GHz active.
     
  22. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #22
    excuse my lack of knowledge..access points are the airport devices and hosts are the devices connecting to each airport?
     
  23. techwarrior macrumors 6502

    techwarrior

    Joined:
    Jul 30, 2009
    Location:
    Colorado
    #23
    Sorry, I guess I got ahead of myself... Hosts are any device on a network. A client device like iPhone, or Mac and servers such as routers or access points are all hosts.

    An access point is a host that provides access to wireless network services to other hosts. A router may also be an access point if it has WiFi enabled.

    So, all of the Airport devices are hosts that become access points when Wi-Fi is enabled on them. But, just to confuse matters, your iPhone might become an access point if you enable Hotspot to share your cellular data connection over WiFi to your Mac such as when traveling.

    Lots of terms, useful for network pros who want to bedazzle or befuddle the unwashed masses :)
     
  24. gavcav thread starter macrumors newbie

    gavcav

    Joined:
    Feb 1, 2017
    #24
    yup..i is a dirty one!:)
    --- Post Merged, Feb 7, 2017 ---
    i'm actually finding this very interesting..much to the disdain of my wife!..i've just got the network analyser app and it looks a useful tool for networking, so i'll have a play over the next day or so and certainly get back to you. one thing to mention is i am quite fortunate in that i have no interference from any neighbours as we live down a quiet country lane and are quite isolated...so is it just a case of trying to keep the 'traffic' down on certain channels, and keeping 2.4GHz to a minimum? i do have certain devices such as ring doorbell and stick up security cameras that i use 2.4GHz because of the distances and obstacles involved to get a signal to the devices..so could i use 5GHz transmitting from the extremes within the centre of the house and then connect express's situated on the house extremities (for example the porch and kitchen) transmitting in 2.4GHz for the ring cameras and ring doorbell? plan image attached.. Screen Shot 2017-02-08 at 02.18.54.png
     
  25. BrianBaughn macrumors 601

    BrianBaughn

    Joined:
    Feb 13, 2011
    Location:
    Baltimore, Maryland
    #25
    The distances in your home are small for a wifi network which leads me to wonder if your building contains a high amount of metal framing.

    My experience with Airport Extremes is that they put out a good, strong signal. I tested one against my own Linksys 1900AC router and, at the time, I think the AE did a better job at the location I was configuring.

    We have a two-story home plus basement and attic with 87 year-old wood framing and the Linksys, located near a corner in the attic, reaches everywhere, even the basement on 5GHz.

    Keep spot testing and I'll bet you'll see much stronger signals where there's a "line of sight" between devices.
     

Share This Page