Upgrade options

[davidfra...@gmail.com]

I have a network with a number of MR3020 and one WR842n-v3 flashed with SECN-4. I am considering a number of upgrade and replacement routes:

1. UPGRADE TO SECN-5 AND ADD GL-MT300A

Am I right in concluding that because of the 4mb memory cap on the MR3020 upgrade SECN-5 now includes 802.11 and this is incompatible with the GL-MT300A? So do I keep the SECN-4 firmware and do not upgrade the WR842-v3 either

2 START AGAIN USING THE GL-MT300A

can you confirm that upgrading the WR842n-v3 to SECN-5 and using the GL-MT300A is technically possible?

3.START AGAIN USING THE AR750M

I see that you have some firmware for the AR750M, technically this is exciting but the unit cost is higher; so how many units would I need on a narrow 3 story building of old bricks that absorb radio waves. A tricky question?

Has anyone got any experience that will help with my problems?

David

[T Gillett]

Hi David

One question please: how many mesh nodes are there in total on your network?

You are correct about the 4MB Flash limitation with recent versions of OpenWrt.

Some background:

As you are no doubt aware, the OpenWrt project was forked in 2016 to form the LEDE project after OpenWrt 15.05 (CC) stable release.

The LEDE project produced the LEDE17.01 stable release in 2017.

The two projects then merged and now have produced the OpenWrt 18.06 stable release.

SECN 4 firmware is based on OpenWrt 15.05 (CC) stable release which supports the 4MB Flash devices like the MR3020.

SECN 5 GA1.0 firmware is based on LEDE 17.01 firmware which also supports 4MB Flash devices. However, there is not a lot of room left in the Flash memory space to load packages, so there are some limitations on what facilities can be included.

SECN 5 GA1.1 firmware is based on OpenWrt 18.06 which does not support 4MB Flash devices (in general), and specifically not the MR3020.

In practical terms, MR3020 and similar devices are now limited to running OpenWrt 15.05, or perhaps LEDE 17.01, but ongoing support and updates for these will obviously be limited.

Conversely, newer devices such as the AR300M and AR750 are supported in OpenWrt 18.06 onwards.

Older devices with 8MB Flash (such as the WR842 V1, V2 and V3) are also supported on OpenWrt 18.06, and so are supported on SECN 5.

In the SECN 5 firmware, we made a decision to switch from using AdHoc to Meshpoint wifi interface for the mesh, because ongoing development is focused on Meshpoint, and not AdHoc.

This means that you can't mix SECN 4 and SECN 5 devices on the same mesh.

So if you want to move to SECN 5 to utilise the newer devices, then you have to leave the older 4MB devices behind.

(One caveat - technically you can switch from AdHoc to Meshpoint in SECN 4 by making some changes in the code, which would allow you to operate SECN 4 devices on a mesh with SECN 5 devices. This could extend the life of your MR3020 devices, but they would eventually run out of support on OpenWrt 15.05)

To address your specific options for upgrading your network:

Option 1 - Switch to SECN 5 and replace all devices with newer devices (operating on 2.4GHz).

This is certainly a viable option.

(Personally, my preference is for the AR300M device over the MT300A device because the wifi drivers for the Atheros chipsets seem to be more mature that those for the MT chipsets.)

Option 2 - Use a mix of existing WR842 V3 and some newer devices all running SECN 5.

This is also a viable option as the WR842 is supported in OpenWrt 18.06 and SECN 5

(Caveat - I test WR842 code on V1 devices only as that is all I have. But others have tested V3 devices with SECN firmware previously)

Option 3 - Use AR750 devices to replace all existing devices

(I presume that in this option you would use the 5GHz radio for the mesh and the 2.4GHz radio for AP.)

This is also a viable option, and gives you the advantage of being able to run the Mesh and APs on independent radios in each device instead of having to share the resources of one radio across both functions.

You would have to test the 5GHz radio in your situation to ensure that it will provide good connectivity around the buildings, given the difference in propagation characteristics between 5GHz and 2.4GHz. The AR750 has internal pcb antennas for the 2.4GHz radio, but has a separate 5GHz antenna mounted internally and connected via a standard pigtail connector on the board. This would allow you in theory to fit an external antenna to  improve the signal propagation in your situation (keeping in mind the legal limits on effective radiated power in your jurisdiction).

Of course you can also use the AR750 to provide a 5GHz AP which might be useful in situations where the 2.4GHz band is overcrowded.

I hope this helps. Please don't hesitate to ask further questions if necessary.

Regards

Terry

--
You received this message because you are subscribed to the Google Groups "Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/0589def7-5286-4954-8660-441ab115b980%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[davidfra...@gmail.com]

On Monday, October 22, 2018 at 5:53:16 AM UTC+1, tlg wrote:

Hi David

One question please: how many mesh nodes are there in total on your network?

There are 5 nodes, 4 MR3020 and1 WR842a-v3; all on Txpower 17 and channel 1. I sometimes wonder if these are the correct options whether I could use Txpower 25 and If I am getting interference using one channel?

You are correct about the 4MB Flash limitation with recent versions of OpenWrt.

Some background:

As you are no doubt aware, the OpenWrt project was forked in 2016 to form the LEDE project after OpenWrt 15.05 (CC) stable release.

The LEDE project produced the LEDE17.01 stable release in 2017.

The two projects then merged and now have produced the OpenWrt 18.06 stable release.

SECN 4 firmware is based on OpenWrt 15.05 (CC) stable release which supports the 4MB Flash devices like the MR3020.

SECN 5 GA1.0 firmware is based on LEDE 17.01 firmware which also supports 4MB Flash devices. However, there is not a lot of room left in the Flash memory space to load packages, so there are some limitations on what facilities can be included.

SECN 5 GA1.1 firmware is based on OpenWrt 18.06 which does not support 4MB Flash devices (in general), and specifically not the MR3020.

In practical terms, MR3020 and similar devices are now limited to running OpenWrt 15.05, or perhaps LEDE 17.01, but ongoing support and updates for these will obviously be limited.

Conversely, newer devices such as the AR300M and AR750 are supported in OpenWrt 18.06 onwards.

Older devices with 8MB Flash (such as the WR842 V1, V2 and V3) are also supported on OpenWrt 18.06, and so are supported on SECN 5.

In the SECN 5 firmware, we made a decision to switch from using AdHoc to Meshpoint wifi interface for the mesh, because ongoing development is focused on Meshpoint, and not AdHoc.

This means that you can't mix SECN 4 and SECN 5 devices on the same mesh.

So if you want to move to SECN 5 to utilise the newer devices, then you have to leave the older 4MB devices behind.

(One caveat - technically you can switch from AdHoc to Meshpoint in SECN 4 by making some changes in the code, which would allow you to operate SECN 4 devices on a mesh with SECN 5 devices. This could extend the life of your MR3020 devices, but they would eventually run out of support on OpenWrt 15.05)

You seem to be offering me a partial solution here; I could switch from AdHoc to Meshpoint in SECN 4 by making some changes in the code; upgrade the WR842n-v3 to OpenWrt 18.06 but relying on comparability with v1 code; and add AR300M flashed with SECN-5 18.06 firmware. Sounds easy in theory haha.

How easy is the switch from AdHoc to Meshpoint in SECN 4?

To address your specific options for upgrading your network:

Option 1 - Switch to SECN 5 and replace all devices with newer devices (operating on 2.4GHz).

This is certainly a viable option.

(Personally, my preference is for the AR300M device over the MT300A device because the wifi drivers for the Atheros chipsets seem to be more mature that those for the MT chipsets.)

Option 2 - Use a mix of existing WR842 V3 and some newer devices all running SECN 5.

This is also a viable option as the WR842 is supported in OpenWrt 18.06 and SECN 5

(Caveat - I test WR842 code on V1 devices only as that is all I have. But others have tested V3 devices with SECN firmware previously)

Option 3 - Use AR750 devices to replace all existing devices

(I presume that in this option you would use the 5GHz radio for the mesh and the 2.4GHz radio for AP.)

This is also a viable option, and gives you the advantage of being able to run the Mesh and APs on independent radios in each device instead of having to share the resources of one radio across both functions.

Yes, I had that option in mind. The only unknown is how many I need for good coverage. I have a signal strength app on my phone so perhaps I get three and take some measurements and decide if a 4th is needed.

[T Gillett]

Hi David

I guess I have to ask the question what do you want to achieve overall?

It does not appear that you want to add additional nodes to the network. Is this correct?

Are you experiencing some problems with the network that you want to address by upgrading the software?

If the network is performing adequately, then there may be no real point in upgrading the firmware version, particularly if it means that you have to replace most of the hardware to do so.

Regarding your question about using all devices on the same channel and the TxPower level, it is necessary to run all devices on the mesh network on the same channel so that they can communicate with each other. The key thing to check is whether there are lots of other devices using the same channel, and thus causing wifi congestion.

For the MR3020 device, the TxPower level is set internally at 18dBm (afaik) and so you don't have much room to move there.

Keep in mind that on the 2.4GHz band, there are only three non-overlapping channels (1, 6 and 11). You should choose one of these channels that is least busy in your situation (and discourage your neighbours from using intermediate channels).

On the 5GHz band, you have more channels to choose from, and a lot less general usage currently, and thus less congestion.

For checking the mesh radio coverage (5GHz or otherwise) you can use the information on the Status page, which shows, for a particular node, the signal strength of all the neighboring nodes.

Modifying the SECN 4 code on your devices to use Meshpoint is not a huge job (probably a day's effort to nut out the details and apply to your devices and test) but you have to realise that you then have your own customised code on your devices with whatever future maintenance effort that implies (i.e. you hack it, you own it :-)

Regards

Terry

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/521758da-afc7-40e5-8572-c6658a762777%40googlegroups.com.

[davidfra...@gmail.com]

On Monday, October 22, 2018 at 12:52:30 PM UTC+1, tlg wrote:

Hi David

I guess I have to ask the question what do you want to achieve overall?

There are two problems: one is poor throughput at one end of the building and the second the slowness of video streaming.The first I thought I could address with a AR300m, but the second a total upgrade

It does not appear that you want to add additional nodes to the network. Is this correct?

Yes

Are you experiencing some problems with the network that you want to address by upgrading the software?

If the network is performing adequately, then there may be no real point in upgrading the firmware version, particularly if it means that you have to replace most of the hardware to do so.

Regarding your question about using all devices on the same channel and the TxPower level, it is necessary to run all devices on the mesh network on the same channel so that they can communicate with each other. The key thing to check is whether there are lots of other devices using the same channel, and thus causing wifi congestion.

For the MR3020 device, the TxPower level is set internally at 18dBm (afaik) and so you don't have much room to move there.

Keep in mind that on the 2.4GHz band, there are only three non-overlapping channels (1, 6 and 11). You should choose one of these channels that is least busy in your situation (and discourage your neighbours from using intermediate channels).

On the 5GHz band, you have more channels to choose from, and a lot less general usage currently, and thus less congestion.

For checking the mesh radio coverage (5GHz or otherwise) you can use the information on the Status page, which shows, for a particular node, the signal strength of all the neighboring nodes.

Modifying the SECN 4 code on your devices to use Meshpoint is not a huge job (probably a day's effort to nut out the details and apply to your devices and test) but you have to realise that you then have your own customised code on your devices with whatever future maintenance effort that implies (i.e. you hack it, you own it :-)

I think that I would be prepared to accept one last upgrade of firmware, and that would also give me the flexibility of replacing a failed MR3020 with a MR300M rather than junking the whole setup if one modem packs up.

Regards and thanks

[T Gillett]

Hi David

Given the problems you are experiencing, it is quite likely that just updating the devices and firmware will not solve the problem.

You could be lucky, but there are probably other factors at play.

Probably worth looking into the issues with the current network before making changes.

Once you understand what is causing the problem, then you can decide whether changing the equipment will help.

A repeatable way of measuring end to end performance of the network is a good starting point.

I use two laptops connected to the network nodes being assessed, and scp a large file between them.

The scp utility will give you stats on what the throughput is end to end, very repeatable and independent of what the network is.

Pushing wifi networks through buildings is not easy. It is quite different to an open field situation.

Walls and floors attenuate the GHz signals very quickly, and cause multiple reflections.

If there is no line of sight between nodes, then the wifi signal will struggle to make it very far through the building.

So you might look at how the nodes are positioned and what the signal strength is between nodes.

Look at the SECN Status page from a particular node to see what the signal strength and Tx/Rx modes are to get an idea of whether you have a workable link between nodes.

If you haven't got 20dB or more SNR (signal to noise ratio) on a link then you are unlikely to get good throughput.

Remember also that if there are multiple hops across the mesh for the data to get from A to B, then the data rate is halved at each hop (for single radio nodes) as the resources of the node radio are shared between the two hops it is supporting.

What is the wifi environment like in the building?

A simple phone app like WiFi Analyzer can give you a good picture of the wifi status.

Are there many other Access Points operating other than your network?

Are they on the same or nearby channels to your network?

Regards

Terry

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/9af36232-38c1-4a84-9bd7-4cf0320bd5ef%40googlegroups.com.