Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: 1 [2] 3

Author Topic: Developments in 802.11ac and after  (Read 9245 times)

currytop

  • Reg Member
  • ***
  • Posts: 114
Re: Developments in 802.11ac and after
« Reply #15 on: January 01, 2016, 09:24:05 PM »

The prices of the Ubiquiti real top-end kit at the AAISP shop
    http://aa.net.uk/broadband-accessories.html
are a bit scary. How do they compare on price?

I don't think A&A are too far off on price for the Ubiquiti range. I have used BroadbandBuyer & 4Gon in the past to buy Ubiquiti product. Unless you really need ac radios you'll save a lot by buying Unifi Pro or plain Unifi product which could be upgraded later when prices have stabilised downwards. For well regarded high end ac products from such as Cisco & Ruckus, they are even more expensive than Ubiquiti at the moment. All the ac products seem to be undergoing regular development & fixes currently.

Steve
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Developments in 802.11ac and after
« Reply #16 on: January 01, 2016, 09:34:38 PM »

Are the lower end products upgrade able to support ac? Because I don't have any ac-capable kit st all at the moment, but I soon will have, hopefully, when I get a newer iPad.
Logged

currytop

  • Reg Member
  • ***
  • Posts: 114
Re: Developments in 802.11ac and after
« Reply #17 on: January 01, 2016, 10:16:51 PM »

Are the lower end products upgrade able to support ac?

Not in any practical fashion since they require different hardware. The upgrade path implies replacement with ac capable hardware.

Steve
Logged

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 43603
  • Penguins CAN fly
    • DSLstats
Re: Developments in 802.11ac and after
« Reply #18 on: January 01, 2016, 11:00:45 PM »

There's a vast amount of information and support on the RPi site https://www.raspberrypi.org/
Logged
  Eric

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: Developments in 802.11ac and after
« Reply #19 on: January 12, 2016, 01:22:13 PM »

A little late to this conversation, but I've been trying to learn about similar things for a residential setup.

I think Ronski makes one key point: otherwise unaided, a client will try to cling on to whatever AP it has itself hooked up to. Even if there is another AP with a better signal available. Even if that AP has the same SSID. Roaming can therefore be a problem.

The better commercially-oriented equipment appears to have features that force a device to switch, or make switching invisible, but I haven't worked out details of how they work, nor how well they work - but nothing appears to be perfect. There are components to the WiFi standards that would help, if only they were widely implemented.

It needs to be noted that roaming has another consequence to behaviour: Commercial AP's seem to be designed to work as a team. There is an expectation that, long before a device gets to the extremities of an AP's range, it would have handed over to a new AP instead. Residential AP's expect to work alone, so expect to need to work with devices at the edges of coverage. These expectations can affect design choices - particularly about the highest power configurations can be.

As an example, Ruckus has a low-end line of "xclaim" models that are meant to be simple to use, and targeted at small businesses. The power available on these models seems to be higher than the new models in the Ubiquiti line (AC-Lite and AC-LR) - yes, higher than Ubiquiti's "long range" model.

Perhaps it is only better to resort to decent commercial equipment when you can expect to saturate the space with plenty of APs - in which case, raw power is not an advantage. Otherwise decent residential equipment is needed - but has to be chosen in the knowledge that roaming will not be easy.

There seems to be quite a set of trade-offs when considering issues of power/range/coverage.

  • Aside: An issue I'm seeing with my Xclaim is that I have a choice of channels in the 5GHz band: "anything", or an individual channel. Setting the device to use one of channels 100-140 allows it to use extra power (up to 30dBm is allowed instead of 23 dBm, 1 Watt instead of 200mW), so can achieve better range.

    Unfortunately, these channels also require the AP to monitor for radar, so it can sometimes switch channel on you if it thinks it has detected radar. The AP can suddenly switch from channel 100 ... but almost always defaults back to channel 36 or 40 when it does so. This is one that allows less power, so range is reduced. They are also the channels that the neighbour's VM superhubs also seem to pick - so you are more likely to collide.

Anyway ...

I was originally trying to get my wireless working using the same SSID for 2.4GHz and 5GHz, so the devices could choose which to be on. With an AP that should steer devices to 5GHz, this should work well, right?

Things didn't seem to work out that way - but being named the same made it hard to determine whether a device really was on 2.4GHz or 5GHz. Then I swapped to separate SSID's, and suffered from @ronski's problem: the device would hang on to a dire signal at 2.4GHz, when a nearby 5GHz signal was a much better choice.

In the end, I too went for an app to help the situation, rather similar to Ronski's: "5 wifi", which attempts to switch to a 5 GHz signal if decent.
Logged

c6em

  • Reg Member
  • ***
  • Posts: 504
Re: Developments in 802.11ac and after
« Reply #20 on: January 12, 2016, 02:05:19 PM »

I know of someone with a v. large house that has such a commercial grade setup from Ubiquiti.
I would imagine it was installed by a pro and not a DIY job.
The place is divided up into several wifi zones and there are multiple access points to ensure the whole place is covered - which indeed do talk to each other to ensure there is proper handover and not this hanging on as long a possible.
As I recall all the access points are linked by network cable.  Not sure if they require a separate server running 24/7 somewhere to manage them or whether they self manage unless you need to change the settings.

I'm told it works exactly as one would expect a proper commercial grade setup to work - properly!
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Developments in 802.11ac and after
« Reply #21 on: January 12, 2016, 02:28:28 PM »

Many thanks c6em, very helpful indeed. You answered what was going to be the next question which was whether or not the Ubiquiti WAPs were all connected by Ethernet cable or not.

I think it's definitely going to be Ubiquiti, but I will wait until I get some AC clients and not go for it before then, nor go for lower end devices.
Logged

currytop

  • Reg Member
  • ***
  • Posts: 114
Re: Developments in 802.11ac and after
« Reply #22 on: January 12, 2016, 10:09:59 PM »

The Ubiquiti AP products use power over Ethernet to connect both data & power. It's much easier to use a POE Ethernet switch to feed power to each of their AP to avoid lots of individual PSUs.

A successful office or large house installation needs careful planning for best performance. Definitely not the most power, but careful selection of non-overlapping channels manually rather than auto set, and careful positioning based on a RF site survey.

The Ubiquiti products are designed to give good average performance to many users at once rather than maximum technically possible throughput to a single user. For this reason they are often criticised when compared to say some Ruckus products when using an unrealistic scenario of a single user trying to maximise throughput. Ubiquiti have looked at this to see if they can tweak performance for this as a special case, but it isn't a design priority.

They have also had some issues with the transparent roaming they advertised on the AC product. That didn't work properly last year, I'm not sure if it has been resolved now.

They are also configured using a management application running on a server. That has to be running when a new AP is added. It could be removed after that but I think it needs to be there if the AP is rebooted or power cycled.
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Developments in 802.11ac and after
« Reply #23 on: January 12, 2016, 10:12:24 PM »

Duly warned.
Logged

currytop

  • Reg Member
  • ***
  • Posts: 114
Re: Developments in 802.11ac and after
« Reply #24 on: January 12, 2016, 10:21:54 PM »

Duly warned.

I didn't mean to warn against Ubiquiti, far from it. In terms of value for money it's hard to beat. But don't get carried away by peak advertised AC speed and expect to achieve anywhere near it for a single device.
Logged

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: Developments in 802.11ac and after
« Reply #25 on: January 12, 2016, 10:47:57 PM »

I meant about its orientation towards multi-user office environments.
Logged

currytop

  • Reg Member
  • ***
  • Posts: 114
Re: Developments in 802.11ac and after
« Reply #26 on: January 12, 2016, 11:10:37 PM »

I meant about its orientation towards multi-user office environments.

Well that's certainly true. The central management of multiple distributed APs is one of its greatest strengths.
Logged

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: Developments in 802.11ac and after
« Reply #27 on: January 13, 2016, 10:51:30 AM »

You answered what was going to be the next question which was whether or not the Ubiquiti WAPs were all connected by Ethernet cable or not.

One key point that may, or may not, be obvious is that the Ubiquiti's are connected by Ethernet cable back to a switch, or a set of interconnected switches, rather than being connected to each other.

Some of the higher-end models do have a secondary ethernet port which is bridged to the primary. That allows another device, which could be another AP, to be daisy-chained. However, the power derived from PoE on the primary port isn't passed on to the secondary - so a new power injector would have to be put onto that line.

The Ubiquiti AP products use power over Ethernet to connect both data & power. It's much easier to use a POE Ethernet switch to feed power to each of their AP to avoid lots of individual PSUs.

Some caution is needed here, as there is a variety of PoE standards.

The lower-end Ubiquiti products (as well as my Ruckus Xclaim) need 24V "passive PoE", while mid-range APs need 802.3af PoE (48V, 15W), and higher-end models need 802.3at PoE+ (48V, 25W).

I was considering a new switch (I'll post about that separately) which supported PoE, and can find ones that document support for 802.3af and at, but specification of support for "24V passive PoE" is scant.

A successful office or large house installation needs careful planning for best performance. Definitely not the most power, but careful selection of non-overlapping channels manually rather than auto set, and careful positioning based on a RF site survey.
Most certainly true - but careful frequency planning can be rendered worthless if the devices detect radar, and shift channel. Does Ubiquiti allow you to configure the fallback channel choice for each AP?

The Ubiquiti products are designed to give good average performance to many users at once rather than maximum technically possible throughput to a single user. For this reason they are often criticised when compared to say some Ruckus products when using an unrealistic scenario of a single user trying to maximise throughput. Ubiquiti have looked at this to see if they can tweak performance for this as a special case, but it isn't a design priority.
Actually - that's a good point. In addition, in an office environment, the aim is often to have more AP's than strictly necessary for coverage - in order to provide plenty of capacity.

It definitely makes for different design.

I think it's definitely going to be Ubiquiti, but I will wait until I get some AC clients and not go for it before then, nor go for lower end devices.

I'm caught similarly, and have almost concluded the same.
Logged

aesmith

  • Kitizen
  • ****
  • Posts: 1216
Re: Developments in 802.11ac and after
« Reply #28 on: January 13, 2016, 08:03:50 PM »

"24V passive PoE" sounds like proprietary device specific power injectors, probably putting power on the non-Ethernet pairs in the way that mid-span injectors do.   I've not seen 24V PoE of any form, or if so I've forgotten.  Given the thin wires involved the developers have always wanted to reduce current, hence the use of 48-50V which is the maximum considered touch safe.   Powering the same device at 24V means twice the current and twice the voltage drop (four times power loss).

As an aside for normal domestic use I've found roaming by default to work perfectly well, just using the same SSID with no AP to AP signalling.  For example at home I have 5GHz in the front of the house and 2.4GHz in the back and in the workshop.   Commercially I think we've only found the need for fancy roaming controls for where customers have wireless handsets on their phone system.
Logged

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: Developments in 802.11ac and after
« Reply #29 on: January 14, 2016, 03:16:58 PM »

It certainly could be proprietary in some way. My best understanding is that it does indeed put power onto unused pairs in the wiring, so look like they can only work on 10/100 fast ethernet links.

It seems unfortunate that Ubiquiti's latest AC access points (the AC-Lite and AC-LR) only support 24V passive PoE, while PoE switches are becoming more available/affordable, but seem oriented at 802.3af.

However, the AC-Lite says it supports gigabit - so perhaps there's a "passive PoE" variant that works with live pairs too.

I *was* contemplating a Netgear GS108PE switch, with 8 ports, 4 of which support 802.3af PoE. It would work with my existing Ruckus Xclaim Xi-2. But it wouldn't work if I swapped to Ubiquiti's AC-Lite devices. My latest searches suggest I would need one of these devices as well: A Ubiquiti PoE Converter.
Logged
Pages: 1 [2] 3
 

anything