Link Aggregation and LACP
HTML-код
- Опубликовано: 19 авг 2024
- I made this video to be a good reference guide for the technologies of LACP and Link Aggregation, two concepts that are often confused. They're related, but separate.
Often when people say LACP, they really mean Link Aggregation. This video shows which one does which.
Fantastic explanation! Thanks for putting the time and effort that obviously went into creating this. Cheers
Definitive video on the topic. Essential viewing.
That was fantastic. Cleared up a lot of misinformation rattling around in my head.
This is by far the best explanation of LAG and LACP i have ever heard. Well done. Confusion cleared. I already knew that LACP is called dynamic LAG because it can discover neighbor switch by sending and receiving LACP PDUs. But i thought static LAG was useless until i watched your video and realized that most of the heavy lifting and operation is done by static LAG and LACP is just an extension to the LAG.
Nice one. Clear explanation. Also fixing confusing terms lol
Really good video. I've configured LAGs many times but to be honest didn't know the details as explained here.
An extremely nice and informative video. I had this confusion about LAG's and LACP. Thanks for correcting me.
Clear introduction thank you!
100 gbit coffee maker, heh. Thank you for this; it is useful without going way into the weeds.
Awesome... Cleared up a bunch of wrong assumptions I was making... Thank you..!
this is a gem❤ , wow what an explanation,thanks for creating this video
Fantastic video! Thank you VERY much for the hard work!
Aaaa. Thank is great! ❤🎉😊 Fantastic!
Great video. Thanks for the simple explanation
Wonderful explanation 👏
Thank you so much for explaining is thoroughly.
I don't know if this is the right place to ask, but please bear with me if you could.
I work with a sound engineer who used Waves Soundgrid protocol for his sound system. Each device in this protocol can be connected via ethernet network, thus the implementation of network switches.
Recently he thought that his network connection, especially from the front of house (mixer control) to stage could use redundancy/traffic regulation via the use of LAG, thus comes the swap from unmanaged switches to managed switches. But the thing is, the switches we bought are only capable of creating static LAG, we then tested that when one of the trunked ports in the same LAG starts to drop in speed (from 1GB to 100MB) AND one of the normal speed Gigabit port is elaborately disconnected; the connection between stage and front of house seems to drop out as well, because a LAG require all ports to run at the same speed and duplex.
This issue can be seen when there are:
- 2 ports within the same LAG (If one drops in speed, the connection ceases)
- 3 ports within the same LAG (if one drops in speed, and the other one got disconnected, the connection ceases)
The issue becomes non-existent with 4+ ports in the same LAG.
The problem is that, it is almost impossible to request every music venue's suppliers to provide us with 4 80-meter ethernet cables from the front of house to stage, because they have to be shared with and cycled between other bands. We could bring our own, but currently as the budget and man power allow; we could only bring 2. We need to be sure that the LAG would still be up and undisturbed during the whole show or the result could be disastrous. The LAG already did traffic regulation pretty well, seeing that there are almost little to no errors packets on the Soundgrid side, but the redundancy is still very unsafe.
Questions:
1. Is there a specific cause to speed drops in the port/cable and how can we prevent this? Because the drop would cause problems for the whole LAG
2. Is there a workaround for this? Or do we need switches that are capable of implementing LACP on LAGs to make them dynamic?
Have you come to a conclusion?
Great video. Before this I would have considered LACP and Link Aggregation the same. I replaced my homelab switch with a L2+ 2.5Gb TP-Link TL-SG3428X-M2 which appeared to have all the features of my 'training' switch which was a very old Cisco. LACP is UP between the switch and NAS but I'm not getting Link Aggregation like my Cisco did which showed 2000rate vs TPlink showing 1000M on the Port Channel. I'll put in a ticket with TP-Link because I'm confused what else to try.
awesome presentation man.
Great explanation !!!
Thanks for the great explanation.
very conceptual explanation
Another key point of Link Aggregation is that while the available bandwidth is increased, the throughput of a flow still remains the same as the packet belonging to the same flow always gets forwarded from the same interface.
Can't we send the flow across different member links? Is it possible to do that?
Good call Tony. Caught myself saying LACP today and I meant Link Aggregation.
Fantastic content!
Ty very much!
Great video! Thanks a lot 😎
Thanks for sharing great video
Great video
Really enjoyed this video and your teaching style. Are there plans for more videos or do you have content elsewhere on the interwebs?
Hi Tony, is it possible to have this slide deck for refresh any time. If yes, please share where I can download the info.
Nice
can you link to vmware docs about lacp?
wondering how it hashes non-IP traffic
wondering how it behaves if an lacp upstream switch stops communicating lacp, does port drop into Independant mode?
I don't know what it does if its set to route based on IP hash and there's no IP header. If you had a lot of non-IP traffic it's probably best to MAC hash or mac pinning.
Hi, thanks for the video. What I dont' get is this. Can a switch that just supports LAG work with another switch that supports LAG with LACP?
Yeah they could both configure a non-LACP LAG, sometimes called a static LAG
@@shadeland thank you
Very detaik
Link aggregation group
?
Great explanation!!