I'm new to the world of cloud AWS, but was able to quickly build a linux ubuntu instance on AWS lightsail and complile the jamulus package. Compiled first time, no issues and server starts fine. ...the first month use is also free, then its $3.50 a month for the basic package.
After setting the above up, i read that AWS also does a "free tier" EC2 instance but with limitations.
Question now is, how much data does jamulus consume with 4 bandmates connecting over AWS. e.g. would the EC2 free tier quota support at least 2 hours of jamming per week or is it best to just stay with the lightsail? Any advice welcomed. Thanks
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Bottom line is that the free tier instances from AWS (whether Windows or linux) have ample bandwidth for any of your needs and certainly much more than jamulus is able to service. So whether your group has 5, 10, 20 or even more it really isn't going to load up EC2. Furthermore the vCPU load is pretty small too. Jamulus is quite a lightweight application.
I use a number of Windows EC2 free tier instances with Jamulus and have never hit anywhere near a limit. Much better to spend time optimising the client end, and how they connect to the Internet.
The EC2 free tier usage limit equates to leaving your server on 24x365 - I never bother to shut mine down. Although I do a precautionary re-start whenever I update Jamulus. So 2hr/week isn't going to be an issue.
Last edit: Andrew S 2020-10-20
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Perfect Andrew! So technically, i could just leave the EC2 free tier on, using the t2.micro . There's less than 750 hours in a month (which is their limit) so all is good. Have i overlooked anything else in terms of bandwidth costs?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes - if you work it out there's a very slight mis-match between 750 hr/mth and each month of the year (some are very slightly longer) but it adds up to peanuts. I tried other free tiers e.g. from Microsoft and Google, and Windows vs linux, but found no difference in performance. And they're all free - so no difference in price :)
I have set up Jamulus for some private organisations, and I specified a slightly higher tier - which came with slightly better bandwidth. But this was more as a "belt and braces" approach to get the best possible performance - for AWS the higher tiers have slightly more advanced networking capabilities to play with.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It does state: A typical jam might have 4 people, for which you would need 200Kbps * 4 = 800Kbs (0.8Mbps) up and down. So if you have a 10Mbits down and 1Mbps up broadband connection
My question was aimed to our fellow jamulus cloud peeps that have actually setup an AWS instance and can advise on cost /time / quality, pros & cons using either service. thanks
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I fogot to say you have to copy the link into your browser. The comma in it means that the normal browser parsing does not work.
In any case, the bandwidth depends on the client setups. You can not determine it youself unless you force the clients to all use a certain setup. You can, however, calculate the maximum usage possible using the information on the linked page.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've been using the EC2 t2micro instance for a while now with the band without issue but for the past 2 weeks, when we connect to the server we are experiencing 'warbling' or choppy audio. No hardware has changed our side. Have restarted the ec2 instance, apt update also. Any ideas what may be causing this issue?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've noticed the same recently, also with the free t2micro. Last night we maxed out CPU at 100% for the Jamulus process with about 12 people. Don't know why. Seems like every clients takes up about 10% cpu on t2micro. It is a one core CPU.
I also noticed that the central server somehow sounds better. Tried it with just me on it. Any idea why?
Last edit: hk1020 2021-02-27
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Have you been leaving your instance on all the time, or switching it off when you're not using it? If you're switching it off you may be running out of CPU credit to support burst requirements, depending on how long your sessions are. The t2 instance types acrue credit when they are idle, which is then used to support high CPU when required, until your credit has run out. So there might be some momentary spike in CPU activity from another task, for example, and if you've run out of CPU credit the jamulus task might get throttled back.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Interesting, didn't know that. I am still on the free tier and hadn't switched off the instance. Jamulus was the only thing running and also the only process at 100%. l get regular warnings at the end of the month that l've used 85% of my free 750 hours. This must be connect time as 750 hours is about 24 * 30 and Jamulus hasn't been running at 100% for the whole month.
Hi Andrew, I have been leaving the t2.micro on 24/7. Just never clicked until now that the service performance could be affected by leaving it on! just tested a t3.micro but performance wasn't that much better to t2.micro. Still getting warbling when the whole band connects! Be good to know from benchmark test which AWS instance gives best performance for jamulus even if we have to pay a few £! for the privilege. Will be used for less than 5 hours a week also.
👍
1
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm new to the world of cloud AWS, but was able to quickly build a linux ubuntu instance on AWS lightsail and complile the jamulus package. Compiled first time, no issues and server starts fine. ...the first month use is also free, then its $3.50 a month for the basic package.
After setting the above up, i read that AWS also does a "free tier" EC2 instance but with limitations.
Question now is, how much data does jamulus consume with 4 bandmates connecting over AWS. e.g. would the EC2 free tier quota support at least 2 hours of jamming per week or is it best to just stay with the lightsail? Any advice welcomed. Thanks
Bottom line is that the free tier instances from AWS (whether Windows or linux) have ample bandwidth for any of your needs and certainly much more than jamulus is able to service. So whether your group has 5, 10, 20 or even more it really isn't going to load up EC2. Furthermore the vCPU load is pretty small too. Jamulus is quite a lightweight application.
I use a number of Windows EC2 free tier instances with Jamulus and have never hit anywhere near a limit. Much better to spend time optimising the client end, and how they connect to the Internet.
The EC2 free tier usage limit equates to leaving your server on 24x365 - I never bother to shut mine down. Although I do a precautionary re-start whenever I update Jamulus. So 2hr/week isn't going to be an issue.
Last edit: Andrew S 2020-10-20
Perfect Andrew! So technically, i could just leave the EC2 free tier on, using the t2.micro . There's less than 750 hours in a month (which is their limit) so all is good. Have i overlooked anything else in terms of bandwidth costs?
Yes - if you work it out there's a very slight mis-match between 750 hr/mth and each month of the year (some are very slightly longer) but it adds up to peanuts. I tried other free tiers e.g. from Microsoft and Google, and Windows vs linux, but found no difference in performance. And they're all free - so no difference in price :)
I have set up Jamulus for some private organisations, and I specified a slightly higher tier - which came with slightly better bandwidth. But this was more as a "belt and braces" approach to get the best possible performance - for AWS the higher tiers have slightly more advanced networking capabilities to play with.
Please see the documentation:
https://github.com/corrados/jamulus/wiki/Network-Requirements
Last edit: Gilgongo 2020-10-20
Hi Don, the link above doesn't take me a 'quality' section but i read the wiki. https://github.com/corrados/jamulus/wiki/Running-a-Server
It does state: A typical jam might have 4 people, for which you would need 200Kbps * 4 = 800Kbs (0.8Mbps) up and down. So if you have a 10Mbits down and 1Mbps up broadband connection
My question was aimed to our fellow jamulus cloud peeps that have actually setup an AWS instance and can advise on cost /time / quality, pros & cons using either service. thanks
I fogot to say you have to copy the link into your browser. The comma in it means that the normal browser parsing does not work.
In any case, the bandwidth depends on the client setups. You can not determine it youself unless you force the clients to all use a certain setup. You can, however, calculate the maximum usage possible using the information on the linked page.
Fixed that rogue comma now.
Excellent !
I've been using the EC2 t2micro instance for a while now with the band without issue but for the past 2 weeks, when we connect to the server we are experiencing 'warbling' or choppy audio. No hardware has changed our side. Have restarted the ec2 instance, apt update also. Any ideas what may be causing this issue?
I've noticed the same recently, also with the free t2micro. Last night we maxed out CPU at 100% for the Jamulus process with about 12 people. Don't know why. Seems like every clients takes up about 10% cpu on t2micro. It is a one core CPU.
I also noticed that the central server somehow sounds better. Tried it with just me on it. Any idea why?
Last edit: hk1020 2021-02-27
Have you been leaving your instance on all the time, or switching it off when you're not using it? If you're switching it off you may be running out of CPU credit to support burst requirements, depending on how long your sessions are. The t2 instance types acrue credit when they are idle, which is then used to support high CPU when required, until your credit has run out. So there might be some momentary spike in CPU activity from another task, for example, and if you've run out of CPU credit the jamulus task might get throttled back.
Interesting, didn't know that. I am still on the free tier and hadn't switched off the instance. Jamulus was the only thing running and also the only process at 100%. l get regular warnings at the end of the month that l've used 85% of my free 750 hours. This must be connect time as 750 hours is about 24 * 30 and Jamulus hasn't been running at 100% for the whole month.
Last edit: hk1020 2021-02-28
Hi Andrew, I have been leaving the t2.micro on 24/7. Just never clicked until now that the service performance could be affected by leaving it on! just tested a t3.micro but performance wasn't that much better to t2.micro. Still getting warbling when the whole band connects! Be good to know from benchmark test which AWS instance gives best performance for jamulus even if we have to pay a few £! for the privilege. Will be used for less than 5 hours a week also.