Micro site scenario

I have recently set up a micro site with the following configuration.

PMP450 AP-->PMP450 SM (UnCapped) --> unmanaged switch--> 2 ePMP 1000 APs ( 1 north, 1 south) -- the south ePMP AP is connectorized to a 90deg antenna and has 4 customers on it. The north AP is basically PTP with one customer.

We use PPPOE to authenticate customers; the PPPOE username and password is in the customers’ router. QoS is done by the ePMP. All customers limited to 3Mbps down 1Mbps up. ePMPs are not synced, but north and south are on different freqs and separated by at least 20 vertical feet. Max Customer distance is 4/10 mile, closest is 1/10 mile.

The site is up and running, kind of.—1 customer is intermittently getting “Limited Internet Access” and others are complaining of slow connectivity. I get 3Mbps when I do speed tests on site, but I haven’t been there during peak hours either. And, lets face it- 3Meg aint that fast anymore. 

I guess my question is--- should this even be working? Can the 450SM handle the multiple PPPOE customers running through it? Are there settings in the 450SM or AP that I need to change for this set up?  Do I need a managed switch - VLANS?

I would have liked to PTP450 to this site, but my ring of PMPs is the last thing the tower owner will allow.

 The 450 SM is the only one on the 450 AP.  ( I have a ring of 6-450 APs, due to the terrain that #4 is pointing at I have never been able to pick up one customer with it.  So when I finally found a customer I could hit (after I put up a 60 ft antenna) I gave him free service to pick up houses below him.  There is potential for other customers if this works-----or should I stop now?

Thank you for any help or ideas.

1 Like

If QoS is done by the customer's ePMP SM... what's PPoE used for? When you do a linkeval on the PMP450 SM that's feeding this repeater site, what sort of throughput do you get?

I almost think you should try something a bit more simple and start there. Use PMP450 radios in bridge mode, install a small managed switch at the repeater site. Get rid of PPoE unless you absolutly need it.

1 Like

Its getting 46Mbps Down 15Mbps up  Currently transmitting at:VC 18 Rate 8X/4X MIMO-B

As far as PPPOE-- I have always used it.  The pppoe is provided by my affiliate ISP to monitor usage, provide 24/7 customer service and all those good things that they do. Could I put the pppoe in the Receiving SM, have it authenticate and pull the IP from the DHCP, then go through a managed switch to the ePMP AP ->customers? This may be a simple explanation (missing some steps) because I don’t fully understand how to do this.

---I can have a static IP assigned to the 450SM---

I have done this several places with 450, 430, even PMP100.  I assume there is no PPPoE going on at the repeater site, just bridged (although having something like a small Mikrotik there would help with troubleshooting).

Are you sure your SM and AP at the repeater site aren't interfering with each other?

I suspect PPPoE is a red herring in this scenario.

1 Like

The 450SM is 5.8Ghz and the ePMPs are 2.4Ghz. I did not separate them (as shown in the pic). Should I separate them by 6 feet? The same as a site with multiple 450 SMs?

No, you shouldn't need to separate them as they're in completely different bands. So going back to what I said... I really do think at least for testing purposes you should simplify things and test each stage in the link... going all the way to the client and doing bandwidth tests NOT using PPoE and seeing what you get, and then using PPoE.

2 Likes

The link tests pretty much mirror internet speed tests on ookla test sites. I can’t remember what the ePMP did with no QoS. Link tests and Internet speed tests were spot on when QoS was enabled. ( 3 QoS groups 3,7 and 10mbps down load speeds). All tests hit right at or just above the limits. Even from the customer router.
The 450 gets 46-50mbps down load consistently. Neither radio is dropping. Up times show they all stay up during a failure, no added reg or re-reg counts.
ps. I can’t connect to Internet without PPPoE.

I was able to cause the link to fail several times.

The ePMP shows a signal strength hovering around -46.

Link test averages 80Mbps down (MIR profile 0)

To test I established a pppoe link with broadband connection from my laptop (no router)

I started a continuouse ping of the ePMP AP.

I watched the pings and the signal strength for at least 5 minutes without a single missed ping or signal drop.

The instant I refreshed a youtube page and a video began to download--

the link failed, eAlign went to "scanning" --- pings failed.

when the pppoe session was manualy ended---

pings to AP returned and after a few seconds the signal returned to -46.

This test was repeated many times.

I replaced the ePMP SM with a new radio.--same results

I tried setting MIR to all three Limits (3,7,15mbps down) (75/25% split)

I also lowered AP power output to the lowest output that still allowed all customers to see a -75 signal.

          lowered SM outputs also.

All test results were the same.

Sometimes I could cause the failure by just beginning a speed test.

I dont think this is a link issue, the radios a clear line of site, under 4k feet. Something is causing it to fail or reset when it passes traffic.

Could I have an incorrect setting in the ePMPs?

If any one has any ideas or has experienced this I would greatly appriciate some help.

Is there a bluprint of the ideal settings for this setup? 

thanks

upgraded all the ePMPs to v2.6.5-RC5.tar.gz

Issue seems to be resolved.

2 Likes

As of today we have had zero drops from the ePMP. I even opened up the customers QOS to 7Mbps to "make it right" for the troubles they have been having. I actualy did it to stress the system. I have been told that the kids are pushing it hard with NetFlix and Xbox,but so far no troubles at all.

PS. Two of the 5 customers have already signed up to keep the higher packages. --This may be a new marketing strategy (for me anyway). Any one else ever try giving customers a "free upgrade" for a week or weekend?

2 Likes

AJones wrote:  Any one else ever try giving customers a "free upgrade" for a week or weekend?

Yea - and then they think it's broken when you throttle them back to normal. :)  So then you explain to them that - "we turned it up temporarily as a (thank you/bonus/demo/mistake/test/reward/example) to our customers" and they are like - "No, I think that's the normal speed it should be all the time, and you're actually (penalizing/sanctioning/impeding/restricting/punishing/reducing) it now".  ;)