Here is the scenario: We have installed remote power monitors on a primary feed (13,200 VAC) to a major hospital complex. It is measuring overall usage, demand, momentary usage (3 seconds for graphing purposes), peaks and surges. The idea is that it may be remotely monitored. As it happens, it is just a bit too far from the nearest node to be reliable.

The installer purchased a range-extender (Cisco). The range extender will connect to the local WiFi just fine. Various "things" will connect to the range extender, just fine. Those things, when connected directly to a node operate just fine. Those things when attempting to connect through the range extender act blocked - "No Internet Connection".

The system is unsecured, all security, firewalls and so forth must be software based and resident on the individual computers.

Is there an easy or obvious solution? I am not the installer, I did not conceive or design the system, I am merely asking on his behalf based on what he is reporting. However, I will be one of the end-users.

Thanks in advance!

Peter Wieck
Melrose Park, PA

On Thu, 17 May 2018 13:34:38 -0700 (PDT), "
wrote:

The installer purchased a range-extender (Cisco). The range extender will connect
to the local WiFi just fine. Various "things" will connect to the range extender,
just fine. Those things, when connected directly to a node operate just fine.
Those things when attempting to connect through the range extender act
blocked - "No Internet Connection".

Wi-Fi range extenders, also known as store and forward repeaters, and
range extenders, can be made to work, but only under limited
conditions and topologies. My comments also apply to mesh networks
and WDS networks, which are range extenders with the addition of an IP
routing protocol.

I've had quite a bit of experience ripping such abominations out and
replacing them with other solutions. I can explain in detail, but I'm
busy right now. Bug me in a few days.

I once gave a talk on the subject of repeaters. Here are my
disorganized notes on the topic:
http://802.11junk.com/jeffl/FLUG-talk-2015-02-28/
To illustrate the effect, I took a simple 2.4 wireless link, and ran
iPerf (actually Jperf) to measure the maximum throughput through the
wireless link. Slides for 802.11g (2.4Ghz) without a range extender:
http://802.11junk.com/jeffl/FLUG-talk-2015-02-28/802.11gn%20direct.jpg
The average throughput was 42 to 61 Mbits/sec average.
I then powered on a pre-configured Linksys RE-2000 range extender and
got this mess:
http://802.11junk.com/jeffl/FLUG-talk-2015-02-28/802.11gn%20through%20Netgear%20repeater.jpg
with an average throughput of zero to 24 Mbits/sec. If I turn off the
range extender, and the speeds return to normal. You can easily
reproduce the effect with a simple wireless router or access point, a
client computah, and Jperf:
https://sourceforge.net/projects/iperf/files/jperf/jperf%202.0.0/
Be careful NOT to mix version 2 and 3 of iPerf and Jperf, which are
incompatible.

If you don't want to setup iPerf or Jperf, then just use ping. Without
the repeater, just ping the local access point continously using some
kind of ping that displays sequence numbers such as hrping:
http://www.cfos.de/en/ping/ping.htm
Ping away merrily and hopefully, you'll only have a few dropped
packets. Then, turn on the repeater and ping away merrily again. If
I'm right, you'll see a dramatic increase in dropped packets. Be sure
to try the -M option (tx/rx times).

Is there an easy or obvious solution?

Nope. The obvious start point is to get ride of generic store and
forward repeaters. If you must use repeaters, build a proper mesh
network, with proper routing, not something that just repeats
EVERYTHING that it hears, which is little better than a jammer.
Because there are to paths between the source and designation of every
packed (direct and through the repeater), you're going to have
collisions no matter how you build it. It's those collisions that are
preventing reliable connections between devices.

I am not the installer, I did not conceive or design the system,
I am merely asking on his behalf based on what he is reporting.
However, I will be one of the end-users.

The right answer is either large number of wireless access points,
connected together with a switch, and not using an in-band backhaul.
Take a look at how hotels and such handle a large number of access
points and clients.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On 5/17/2018 1:34 PM, wrote:
Here is the scenario: We have installed remote power monitors on a primary feed (13,200 VAC) to a major hospital complex. It is measuring overall usage, demand, momentary usage (3 seconds for graphing purposes), peaks and surges. The idea is that it may be remotely monitored. As it happens, it is just a bit too far from the nearest node to be reliable.

The installer purchased a range-extender (Cisco). The range extender will connect to the local WiFi just fine. Various "things" will connect to the range extender, just fine. Those things, when connected directly to a node operate just fine. Those things when attempting to connect through the range extender act blocked - "No Internet Connection".

The system is unsecured, all security, firewalls and so forth must be software based and resident on the individual computers.

Is there an easy or obvious solution? I am not the installer, I did not conceive or design the system, I am merely asking on his behalf based on what he is reporting. However, I will be one of the end-users.

Thanks in advance!

Peter Wieck
Melrose Park, PA

What's the make and model of the RE?

Think of a PLC solution.

a écritÂ*:
Here is the scenario: We have installed remote power monitors on a primary feed (13,200 VAC) to a major hospital complex. It is measuring overall usage, demand, momentary usage (3 seconds for graphing purposes), peaks and surges. The idea is that it may be remotely monitored. As it happens, it is just a bit too far from the nearest node to be reliable.

The installer purchased a range-extender (Cisco). The range extender will connect to the local WiFi just fine. Various "things" will connect to the range extender, just fine. Those things, when connected directly to a node operate just fine. Those things when attempting to connect through the range extender act blocked - "No Internet Connection".

The system is unsecured, all security, firewalls and so forth must be software based and resident on the individual computers.

Is there an easy or obvious solution? I am not the installer, I did not conceive or design the system, I am merely asking on his behalf based on what he is reporting. However, I will be one of the end-users.

Thanks in advance!

Peter Wieck
Melrose Park, PA

On 05/17/2018 04:34 PM, wrote:
Here is the scenario: We have installed remote power monitors on a primary feed (13,200 VAC) to a major hospital complex. It is measuring overall usage, demand, momentary usage (3 seconds for graphing purposes), peaks and surges. The idea is that it may be remotely monitored. As it happens, it is just a bit too far from the nearest node to be reliable.


Wow, I've never observed (above ground) anything like on a medium
voltage (MV) AC distribution line. Just MV-LV distribution
transformers, disconnect switches, circuit breakers and power factor
correction caps. (We use 34.5/19.9 kV in my area). Extremely dangerous
if mishandled. Sincerely,

--
J. B. Wood e-mail:

On Friday, May 18, 2018 at 2:35:30 AM UTC-4, Mike S wrote:

What's the make and model of the RE?

Cisco-Linksys

White device, two antenna, direct plug-in.

Beyond which I do not know.

Peter Wieck
Melrose Park, PA

On Fri, 18 May 2018 09:15:42 -0700 (PDT), "
wrote:

On Friday, May 18, 2018 at 2:35:30 AM UTC-4, Mike S wrote:
What's the make and model of the RE?

Cisco-Linksys
White device, two antenna, direct plug-in.
Beyond which I do not know.

Peter Wieck
Melrose Park, PA

My guess(tm) would be Linksys RE6300, RE6400, or RE6700:
https://www.linksys.com/us/p/P-RE6300/
https://www.linksys.com/us/p/P-RE6400/
https://www.linksys.com/us/p/P-RE6700/
The problem with a wall mounted solution is that the antennas are in a
terrible location for best coverage. They should be up as high as
possible and away from any metal objects. These repeaters are made
for installation convenience, not performance.

Incidentally, if they say Cisco on the box, they're at least 3 years
old. Belkin bought Linksys from Cisco in about 2013. The purchase of
Belkin by Foxconn is awaiting regulatory approval.

Someone mentioned power line networking. In a sub-station, you want
as much air-gap isolation from the lines as possible. That leaves
wireless, FSO (free space optics), and fiberoptics as the available
options.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Friday, May 18, 2018 at 12:39:57 PM UTC-4, Jeff Liebermann wrote:

The problem with a wall mounted solution is that the antennas are in a
terrible location for best coverage. They should be up as high as
possible and away from any metal objects. These repeaters are made
for installation convenience, not performance.

Incidentally, if they say Cisco on the box, they're at least 3 years
old. Belkin bought Linksys from Cisco in about 2013. The purchase of
Belkin by Foxconn is awaiting regulatory approval.

Someone mentioned power line networking. In a sub-station, you want
as much air-gap isolation from the lines as possible. That leaves
wireless, FSO (free space optics), and fiberoptics as the available
options.

It was plugged into a high outlet just under the ceiling and line-of-sight, perhaps 30 feet from the node. Keep in mind that this is a hospital basement built in the 1930s before any sort of computing of any nature. The footprint is an entire city block.

We are thinking that the main router has some means of blocking these things out. Is that possible? The nodes are all hard-wired.

On Fri, 18 May 2018 10:37:42 -0700 (PDT), "
wrote:

On Friday, May 18, 2018 at 12:39:57 PM UTC-4, Jeff Liebermann wrote:
It was plugged into a high outlet just under the ceiling and line-of-sight,
perhaps 30 feet from the node.

Altitude will help the signal clear all the ground level obstructions,
such as furniture and people. However, altitude tends to offer a
clear shot at windows, through which much interference from other
services on the same shared frequency band might enter the building.
You might consider moving the repeaters away from any windows with a
view of the city.

The basic problem is that the repeater will be a problem if there are
two paths between the source and destination of a packet. For
example, if the source and designation of a packet can both see each
other, as well as both see a repeater, then two identical packets will
arrive for every packet sent. One direct from the source to the
destination, and one from the source, to the repeater, and to the
destination, after a delay. The delay comes from most Wi-Fi device
being simplex and not full duplex. Simplex is when a device can only
transmit *OR* receive, but not both at the same time. Most Wi-Fi
devices are like that, including the repeater. If the repeater has to
store and forward a packet, there will be a 1 packet length delay
(which varies by packet size) switching from receive to transmit.

You can block the direct path between the source and destination by
making the path go around a corner. The corner obstruction blocks the
direct path. Since both the source and destination can both see the
repeater, the repeated path is not blocked and functional without self
interference. Unfortunately, that doesn't do much when the source is
a portable device, laptop, or smartphone, that can easily be moved
into a situation where there are 2 paths between endpoints.

It appears that you may have missed my first posting up-thread, where
I covered some of the problems with repeaters, mesh networks, and
self-interference. Should I repost it or email it to you?

Keep in mind that this is a hospital basement
built in the 1930s before any sort of computing of any nature. The
footprint is an entire city block.

I don't understand what that has to do with the initial connection
problem for a device.

We are thinking that the main router has some means of blocking
these things out. Is that possible? The nodes are all hard-wired.

Sure. You can filter by MAC address. All, and I do mean *ALL*, Wi-Fi
networking is done on Layer 2, the MAC address layer. The MAC
addresses of the source and destination of every packet are exposed
and not encrypted. There should be a MAC address filter in the router
that lets you block any Wi-Fi device by MAC address. Some of the less
disgusting wireless bridges and repeater also have MAC address
filters. However, I don't see what that buys you for eliminating what
I consider to be a designed in source of interference.

If the "nodes", which I presume means access points, are all wired to
a central router, then the system should have worked as expected,
although apparently with a few dead spots. Remove ALL the repeaters,
add some more hard wired access points in these dead spots, and it
should work.

If ripping out the repeaters doesn't improve the initial connections,
then something else is wrong. I've had to drag an SA (spectrum
analyzer) to the problem location to see what's happening. I've found
noise sources, non Wi-Fi devices, frequency hoppers, powerful inband
carriers, high power cordless phones, and such, none of which show up
on computing based Wi-Fi analyzers. The closest approximations that
might substitutes for a spectrum analyzer are either an SA dongle:
https://www.metageek.com/products/wi-spy/
or an AP (access point) with a built in spectrum analyzer such as the
various Ubiquiti products running Airview:
https://www.youtube.com/watch?v=NAuaSdzHwoo
https://dl.ubnt.com/datasheets/airmax/UBNT_DS_airView.pdf
https://help.ubnt.com/hc/en-us/articles/204950584-airMAX-How-to-Use-airView-to-Find-the-Best-Channel
or various low end portable SA's.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558