How many packets per second????

[SMART]

Hi

we need how many packets per second suport the rockets in the real word not in dreams??????

Some can help, but we need reality answer

regards.

smart

[Ron]

Depends on the size of the packets and what other features are in use.

Why don't you put some effort into doing tests to answer your own questions?

[staslabs]

In my tests less than 15000 pps as sum of both directions.

[ridnet]

Hi

we need how many packets per second suport the rockets in the real word not in dreams??????

Some can help, but we need reality answer

regards.

smart

Answer the following and I will give you an answer

1- What type of application?
2- Do you have other applications, and what?
3- How many equipment will the application register to ( How many MAC )?
4- Will you have a high end router connected to the Rocket at any end?
5- What LAN speed will you use in the Rocket (10 or 100 Mbps )?
6- What modulation will you use in the wireless side of the Rocket ?
7- What speed will you use the wireless side of the Rocket?
8- What is the spectrum analysis telling you?
9- What is the distance of the Link?
10- Is this a PTP or PMP link?

I have other questions, start with them first.

Forget about the 15K PPS, That's impossible under any condition.

[SMART]

answer the following and i will give you an answer

1- what type of application? sales of telephone calls
2- do you have other applications, and what? internet
3- how many equipment will the application register to ( how many mac )? 100 for each rocket
4- will you have a high end router connected to the rocket at any end? yes
5- what lan speed will you use in the rocket (10 or 100 mbps )? 100mbps
6- what modulation will you use in the wireless side of the rocket ? 108 - 108
7- what speed will you use the wireless side of the rocket? 50mbps
8- what is the spectrum analysis telling you? cero interference
9- what is the distance of the link? 8kms
10- is this a ptp or pmp link? ptmp

i have other questions, start with them first.

Forget about the 15k pps, that's impossible under any condition.

perfect mi answers are in your questions in red
thanks...smart

[ridnet]

1- what type of application? sales of telephone calls
Will sales be using SQL, Citrix, Remote Desktop, File Sharing .....

2- do you have other applications, and what? internet
This is a general response,... What will they use the internet for?

3- how many equipment will the application register to ( how many mac )? 100 for each rocket
Is this at each end of the PMP, one Rocket with 100 stations / Phone, devices?
I will assume 100 concurrent calls.

4- will you have a high end router connected to the rocket at any end? yes
What type and model?

5- what lan speed will you use in the rocket (10 or 100 mbps )? 100mbps

6- what modulation will you use in the wireless side of the rocket ? 108 - 108
I will assume it's QAM256, OFDM, MIMO

7- what speed will you use the wireless side of the rocket? 50mbps
I will assume it's 54Mbps

8- what is the spectrum analysis telling you? cero interference
Please forward the trace of the SA.

9- what is the distance of the link? 8kms
I will assume LOS, 80% Fresnel

10- is this a ptp or pmp link? ptmp
This answer makes things impossible.

This is a carrier type solution,
Can you spend a little more money on the radios, like $20K more, per link?

[tcook]

Everyone always dances around answering this question. And always wants more info before answering. How about using the industry standard as a start. List the pps based on each modulation mode unidirectional based on 64byte packets where 8bits=byte. Assuming lab conditions and no other usage. ie., No interference. This is the max PPS that this equipment can achieve. In other words, where does the processor choke.

[staslabs]

In my tests less than 15000 pps as sum of both directions.

10 km link
55 cm dishes
dual polarity
rocket 5m
270-300 mb/s chanel speed
no wireless security

link from ISP network to magistral provider - mixed traffik from many clients.

[SMART]

everyone always dances around answering this question. And always wants more info before answering. How about using the industry standard as a start. List the pps based on each modulation mode unidirectional based on 64byte packets where 8bits=byte. Assuming lab conditions and no other usage. Ie., no interference. This is the max pps that this equipment can achieve. In other words, where does the processor choke.

1- what type of application? Sales of telephone calls
will sales be using sql, citrix, remote desktop, file sharing .....tallk via telephone only

2- do you have other applications, and what? Internet
this is a general response,... What will they use the internet for? browse an navigate

3- how many equipment will the application register to ( how many mac )? 100 for each rocket
is this at each end of the pmp, one rocket with 100 stations / phone, devices? no 100 clients with rocket, and 10 phone lines each
i will assume 100 concurrent calls. this is the question

4- will you have a high end router connected to the rocket at any end? Yes
what type and model? mikrotik rb100

5- what lan speed will you use in the rocket (10 or 100 mbps )? 100mbps

6- what modulation will you use in the wireless side of the rocket ? 108 - 108
i will assume it's qam256, ofdm, mimo yesssss

7- what speed will you use the wireless side of the rocket? 50mbps
i will assume it's 54mbps yes

8- what is the spectrum analysis telling you? Cero interference
please forward the trace of the sa. we dont have anymore

9- what is the distance of the link? 8kms
i will assume los, 80% fresnel yes

10- is this a ptp or pmp link? Ptmp
this answer makes things impossible. the rockets are make for this

this is a carrier type solution,
can you spend a little more money on the radios, like $20k more, per link? the money not is the problem, we need know how many pps for purchase ubnt or not

thank you very much for your help and we waith your answer

smart

[tcook]

It seems these units fall short when passing 64 byte packets. See:

http://www.ubnt.com/forum/showpost.php?p=69927&postcount=7

[sldnkarm]

SMART/All,

There are open source tools you can easily download, install and run to find out the answer to your question. If you don't have access or ways to accomplish this, you have to expect not everyone has either tested this or does not want to hand over information they put their time/sweat into gathering. If you do any sort of equipment evaluations you need to download these tools and start understanding how they work and utilize them for your own evaluations.

With that said, here are some quick results from an earlier vivanet assessor test I completed:
100calls (simul/G.711a) - 4.30MOS - 34.95ms(latency) - .36JitterBufferLoss
150calls (simul/G.711a) - 4.22MOS - 44.34ms(latency) - .55JitterBufferLoss
175calls (simul/G.711a) - 3.42MOS - 139.3ms(latency) - 5.78JitterBufferLoss

This is over a 12mile link with RocketM5/RocketDish:
20Mhz/v5.1/5.830Ghz/90%CCQ on both ends

If you need further information/consultation please contact me via PM.

[tcook]

But what if I am not a customer yet and need this information to make an informed decision. Shouldn't the manufacture have already completed these standardized tests and make them available to potential customers? And to be honest with you, not to make this sound personal, but why would I trust your results? I dont know you or your testing methodology or your test environment. Sure, I can do those same tests. So now I have to buy equipment and test it only to find out that it wont work in my environment. If the data from these standardized tests are made available. I could make an informed decision.

[sldnkarm]

Maybe you could share your experience with the other manufactures you are currently investigating as well. How are they providing this information to you? Do you trust other manufactures results? If so, did you question where the tests were conducted?

Why would you trust anyone who is trying to sell something over someone who is using it in the field and has fully tested it??? That is why I recommend people who are in the business to install/support networking equipment like ubiquiti products, you need to have a certain set of tools to be successful. My advice to you, if you are having issues decided who to trust, is to trust no one and complete the testing yourself.

Also remember that no one given installation is alike and can be supported as such. We are dealing with radio frequencies here. Ubiquiti cannot control the medium everyone uses their equipment on so results WILL vary.

Keep in mind we are talking about sub-$200 devices. If you are trying to compare it to Alvarion (PPS) it is not one for one comparison and you need to decided what it is you are wanting the devices to deliver/support. While I admit these are no where near the PPS performance you will receive with Alvarion, they are quite impressive and are proving to be quite worth the money invested in them.

But why would you trust me?! ;)

[tcook]

A manufacture lists of specs makes them liable to false advertising. But at the same time a lack of stardard specs for a product makes it questionable. PPS in my opinion is pretty standard that far to many manufactures fail to provide without proding. As you see I got my answer from UBNT in the link I posted. Which I will trust over any data anyone else posts since they have more to lose. I am not debating anything about any performace levels about anyones products. Including UBNT's. What I was questioning was for UBNT to list a pretty standard spec. So thank you kind sir for your input, but it did NOT pertain to what or from whom I was asking.

[sldnkarm]

tcook

I was adding additional information for the good of the group. You were not the only one asking questions in this thread and I felt that it would benefit the community to include these results. You comments about how the UBNT products fall short are interesting to me. How did you come to this conclusion? What are some of the other manufactures you are using in your evaluation and what did they 'report' their PPS rate was?

[tcook]

Your kidding right? Do the math based on what UBNT reported in the other thread. They obviously get cpu bound rather quickly. Which is quite ashame for such a good product. But makes it quite ineffecient for someone that needs a backhaul for VOIP. Fortunately thats not me.

[sldnkarm]

Unless you are going to provide information on what you are comparing the UBNT products to, what you say is inaccurate. It's a 400Mhz processor, what are you expecting these devices to do?

[UBNT-Mike.Ford]

Guys,

PPS is also a factor of processing power/overhead. That is why we have conservative numbers over wireless. PPS over wireless takes a LOT of resources. There are things that can significantly increase this (like proper installations, great link CCQ ect) but I will not give you guys higher numbers to have you come back and say "Why did so and so get 30k and I can only get 20k" ect. Better to err on the low side.

Thanks,

Mike

[ridnet]

thank you very much for your help and we waith your answer

smart


Forget about all the questions, last answer changes things.

Mircrotik is not a highend router.
No, the Rockets will not work under this set up you want .

Money is no problem in this project, get 100 pair of PTP license radios, like Harris, Motorola, Cerragon....

Total cost is about $2,000,000.00 with out installation.

Get professional help.

[davey]

SMART/All,

There are open source tools you can easily download, install and run to find out the answer to your question. If you don't have access or ways to accomplish this, you have to expect not everyone has either tested this or does not want to hand over information they put their time/sweat into gathering. If you do any sort of equipment evaluations you need to download these tools and start understanding how they work and utilize them for your own evaluations.

With that said, here are some quick results from an earlier vivanet assessor test I completed:
100calls (simul/G.711a) - 4.30MOS - 34.95ms(latency) - .36JitterBufferLoss
150calls (simul/G.711a) - 4.22MOS - 44.34ms(latency) - .55JitterBufferLoss
175calls (simul/G.711a) - 3.42MOS - 139.3ms(latency) - 5.78JitterBufferLoss

This is over a 12mile link with RocketM5/RocketDish:
20Mhz/v5.1/5.830Ghz/90%CCQ on both ends

If you need further information/consultation please contact me via PM.

Thanks for the data. We are about to deploy Airmax in our network and we will be selling VOIP. This gives us a good idea what is possible.