Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How we solved our office Wi-Fi problems

triplebyte.com

Post reply on HN
Print view View original

231–240 of 252 posts

Re: How we solved our office Wi-Fi problems

#231

Post by inferiorhuman »

kryogen1c wrote: post #215
jdiaz5513 wrote: post #179

Earlier quoted context omitted.

Yup. The network engineer in me would never recommend Ubiquiti for any scenario. Maybe it works for WISP because of the price point; I'd have to be convinced. They run an embedded Mongo DB on their UniFi hardware that (at least in the deployment I've inherited) requires occasional direct interventions[1] to keep running. That's just one example of the many baffling/wrong things they do. I know Mongo gets quite a bit…

My IT MSP solely sells and supports unifi WAPs, and I have no idea what you're talking about. We have dozens of WAPs under management across the city and state, and I've never seen any of the issues you guys are talking about. We don't have issues with database crashing or overheating. The wap controllers are configurable for automatic firmware updating, so I don't know what the issue is with patching. They definitel…

> They definitely support wpa-e/radius because I've configured it and we use it in my my office.

Sure, and problems with RADIUS auth have been well documented. How much of the not seeing issues is simply a matter of users assuming your setup (or their phones or computers) is flakey? Issues with Mongo are also pretty well documented.

https://www.google.com/search?q=ubnt+mongo+issue

> The wap controllers are configurable for automatic firmware updating, so I don't know what the issue is with patching.

The issue with patching is that Ubnt is in a bad position. They're on their own completely to create and issue patches even for stock packages. That's just what happens when you distribute EOL'd software. So, sure, it's not that hard to apply patches* from Ubnt but it's a lot more work for Ubnt to generate these patches — and last I checked they were pretty well behind with the excuse of "there are no known exploits for ABC in the wild".

* Let's not forget the ER-X bootloader for which end users are entirely on their own. Ubnt has a patch but hasn't applied it to the devices they ship. It's just lazy and sloppy.

Re: How we solved our office Wi-Fi problems

#232

Post by snowwrestler »

Grazester wrote: post #218
kryogen1c wrote: post #215

Earlier quoted context omitted.

My IT MSP solely sells and supports unifi WAPs, and I have no idea what you're talking about. We have dozens of WAPs under management across the city and state, and I've never seen any of the issues you guys are talking about. We don't have issues with database crashing or overheating. The wap controllers are configurable for automatic firmware updating, so I don't know what the issue is with patching. They definitel…

I have an office with a few of those and they are not great and do overheat. I am quite surprised you having been exposed to a great number of them never had an issue.

Maybe he hasn't had overheating problems because he is kryogen1c.

(I'll see myself out...)

Re: How we solved our office Wi-Fi problems

#233

Post by jrockway »

Dylan16807 wrote: post #149
jrockway wrote: post #68

Earlier quoted context omitted.

Software has been written that will forcibly disconnect a 5GHz capable client that is found using 2.4GHz. How widely deployed this is, I don't know. Avery did a talk about it (and other things) a couple years ago: https://apenwarr.ca/diary/wifi-data-apenwarr-201602.pdf

What's the point of using one SSID for both frequencies if you make roaming between them impossible?

The AP can make the decision as to whether or not it's possible. It has data about how well the packets are getting through on both bands (as long as the device probes both bands). If it sees that 5GHz is working, but the device is using 2.4GHz, then it can disconnect it from 2.4GHz. But if 5GHz doesn't work, then it can allow the client to stay on 2.4GHz.

Clients do pretty stupid things, which is why the AP is the right place to control this sort of thing. (If you have more than 1 AP, then you have signal strength data from the other APs, and AP A can disconnect the client and force it to use AP B. This seems to work better than letting the client decide on its own.)

Re: How we solved our office Wi-Fi problems

#234

Post by Dylan16807 »

jrockway wrote: post #233
Dylan16807 wrote: post #149

Earlier quoted context omitted.

What's the point of using one SSID for both frequencies if you make roaming between them impossible?

The AP can make the decision as to whether or not it's possible. It has data about how well the packets are getting through on both bands (as long as the device probes both bands). If it sees that 5GHz is working, but the device is using 2.4GHz, then it can disconnect it from 2.4GHz. But if 5GHz doesn't work, then it can allow the client to stay on 2.4GHz. Clients do pretty stupid things, which is why the AP is the r…

Sure, as long as it's using real signal data to make the decision then things are good.

Re: How we solved our office Wi-Fi problems

#235

Post by kryogen1c »

basch wrote: post #226
ShouldBeWorkng wrote: post #224

Earlier quoted context omitted.

You have dozens of devices in what sounds like small locations across your state. I have 11k users online in North America and I would never trust Unifi for any part of my network. They have their place but enterprise is not one of them. SMB is not the same.

I agree fully, its not enterprise grade. Triplebyte is a 30 person office.

All parents and responses to my posts are probably right, and I'm probably wrong!

My experiences aren't wrong, of course, but it's an apples to oranges comparison. My IT MSP is only small to medium sized businesses (as I think all IT MSPs are since Enterprise level environments don't contract out it and hire their own department - this is an educated guess). We do very little Enterprise work and have no single client with more than 300 users.

However, I will defend my position and consumer grade ubiquitis on a (significant) technicality. Higher models beyond what my company sells are not cost effective. Buying higher grade waps at no less than double the price is more expensive than just running cables. If you WANT to make everyone get 200mbps wifi, you will pay for it, but that's a convenience you pay for, not a requirement. UAP-AC-Pros are perfectly capable of handing out up to 50mbps pipes to multiple endpoints without huge (or even medium) upfront capital investments, overheating or DB problems. In my opinion, just run cat6 lines. Itll be cheaper than spending thousands on Enterprise waps for only 30 people.

Re: How we solved our office Wi-Fi problems

#236

Post by jrockway »

Dylan16807 wrote: post #234
jrockway wrote: post #233

Earlier quoted context omitted.

The AP can make the decision as to whether or not it's possible. It has data about how well the packets are getting through on both bands (as long as the device probes both bands). If it sees that 5GHz is working, but the device is using 2.4GHz, then it can disconnect it from 2.4GHz. But if 5GHz doesn't work, then it can allow the client to stay on 2.4GHz. Clients do pretty stupid things, which is why the AP is the r…

Sure, as long as it's using real signal data to make the decision then things are good.

Yup. I can certainly understand the hesitation because it's basically two computer programs both saying "I'm smarter than you! No I'm smarter than you!" which rarely ends well.

But I think since WiFi clients are so varied in their intelligence, the ideal thing is to build the roaming functionality into the AP... because at least you have the ability to fix it once and for all for everyone. Whether or not currently commercially-available hardware does that, I don't know.

Re: How we solved our office Wi-Fi problems

#237

Post by m_eiman »

kraftman wrote: post #102
MBCook wrote: post #56

It’s only sort of passively mentioned in the article but I am AMAZED at the number of people who don’t hardwire everything they can. Obviously phones are out, but why not hardwire every laptop when it’s at the desk? If someone’s using a actual desktop computer like an iMac then what’s the point of Wi-Fi? Clear up the signal space and get a 100% reliable and ultra fast connection.

You could say the opposite I guess. I'm AMAZED at the number of people that bother hardwiring everything they can for no reason when they have no interference issues. Why go to all the effort of hardwiring every laptop at every desk? If someone's using a laptop and moving around the office, why cable every desk when you can have gigabit wifi wherever you are?

Wired:

PING 10.0.3.1 (10.0.3.1): 56 data bytes

64 bytes from 10.0.3.1: icmp_seq=0 ttl=64 time=0.737 ms

64 bytes from 10.0.3.1: icmp_seq=1 ttl=64 time=0.636 ms

64 bytes from 10.0.3.1: icmp_seq=2 ttl=64 time=0.701 ms

64 bytes from 10.0.3.1: icmp_seq=3 ttl=64 time=0.633 ms

...

--- 10.0.3.1 ping statistics ---

8 packets transmitted, 8 packets received, 0.0% packet loss

round-trip min/avg/max/stddev = 0.516/0.644/0.737/0.075 ms

Wifi:

PING 10.0.3.1 (10.0.3.1): 56 data bytes

64 bytes from 10.0.3.1: icmp_seq=0 ttl=64 time=6.713 ms

64 bytes from 10.0.3.1: icmp_seq=1 ttl=64 time=3.508 ms

64 bytes from 10.0.3.1: icmp_seq=2 ttl=64 time=2.425 ms

64 bytes from 10.0.3.1: icmp_seq=3 ttl=64 time=2.127 ms

64 bytes from 10.0.3.1: icmp_seq=4 ttl=64 time=4.057 ms

...

--- 10.0.3.1 ping statistics ---

22 packets transmitted, 22 packets received, 0.0% packet loss

round-trip min/avg/max/stddev = 1.429/3.769/16.722/4.054 ms

Depending on what you're doing, this can make a HUGE difference.

Re: How we solved our office Wi-Fi problems

#238

Post by kraftman »

m_eiman wrote: post #237
kraftman wrote: post #102

Earlier quoted context omitted.

You could say the opposite I guess. I'm AMAZED at the number of people that bother hardwiring everything they can for no reason when they have no interference issues. Why go to all the effort of hardwiring every laptop at every desk? If someone's using a laptop and moving around the office, why cable every desk when you can have gigabit wifi wherever you are?

Wired: PING 10.0.3.1 (10.0.3.1): 56 data bytes 64 bytes from 10.0.3.1: icmp_seq=0 ttl=64 time=0.737 ms 64 bytes from 10.0.3.1: icmp_seq=1 ttl=64 time=0.636 ms 64 bytes from 10.0.3.1: icmp_seq=2 ttl=64 time=0.701 ms 64 bytes from 10.0.3.1: icmp_seq=3 ttl=64 time=0.633 ms ... --- 10.0.3.1 ping statistics --- 8 packets transmitted, 8 packets received, 0.0% packet loss round-trip min/avg/max/stddev = 0.516/0.644/0.737/0.…

I'm interested in what situations 3ms difference (16ms max) makes a difference in an average office environment?

Re: How we solved our office Wi-Fi problems

#239

Post by smolder »

swingline-747 wrote: post #86
linsomniac wrote: post #73

Generally pretty solid advice. I say that as someone who is known for solving tough wireless problems. :-) On the cable termination part: I've (mostly) stopped crimping cables because I've had too many go flaky and don't have 4-5 figure testing equipment. One thing I'll add is that there are ends for solid conductor and stranded, make SURE you have the right ones for the cable you are using. These days I always just…

Been There, Done That, Bought The T-Shirt. Solid core is generally for premise wiring (PVC jacketed in walls, PTFE-jacketed through ducts); stranded is typically for patch cables. If you try like the first place I worked at in the mid-90's trying to put stranded ends on solid-core wire, breaking of tools and unreliable cables will make. There's cheapo Chinese cable tester kits on eBay, AliBaba and Amazon that do a go…

Fluke is one of those used-to-be-good now-abuses-their-reputation companies anyway. They're in a mode of milking consumer trust for what it's worth until it's gone.

Re: How we solved our office Wi-Fi problems

#240

Post by metildaa »

inferiorhuman wrote: post #117
metildaa wrote: post #107

Earlier quoted context omitted.

Note that nearly all of Ubiquiti's offerings are running modified OpenWRT, with a handful running a fork of VyOS which happens to be Debian plus some packages.

> Note that nearly all of Ubiquiti's offerings are running modified OpenWRT, with a handful running a fork of VyOS which happens to be Debian plus some packages. EdgeOS is a Vyatta fork. Has Ubnt abandoned EdgeOS? It's been a few months since I've looked, but the version of Debian that's being used was end-of-lifed on MIPSLE hardware a while back so security updates are entirely incumbent upon Ubnt. There is no upstr…

All of their wireless gear is OpenWRT based, EdgeOS is only used on the switching hardware. Their wireless gear has been mmuch more popular and has significantly more variety than the switches/routers Ubiquiti offers.
Post reply on HN