Timeline for STP frames over VPLS/VPWS
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 7, 2021 at 11:03 | comment | added | Zac67♦ | From the ISP perspective, it might look like the customers are using (more transparent) L2 links but especially larger clients won't ever use those to span large broadcast domains. | |
| Nov 7, 2021 at 11:01 | comment | added | Zac67♦ | @A.A The only advantage a large L2 segment has is that's it's a large L2 segment. Any device can discover anything by broadcast and talk to any other device 'directly'. L2 doesn't scale well (1000+ nodes, anyone?), pollutes your large broadcast domain with said broadcasts, provides little control over what's permitted and what isn't (zoning!), disables many QoS mechanisms, and last but not least, doesn't support redundant or even load-balanced meshing. L2VPN may look simpler and more feasable with just two locations but it quickly becomes an absolute nightmare beyond that. | |
| Nov 7, 2021 at 10:55 | comment | added | A.A | Actually, I didn't get why routing has an advantage over L2VPN. At least L2VPN must have some advantages too... I have heard from a tech person that small customers (like branches of stores or banks) use L3VPN and bigger customers (like ISPs) use L2VPN. @ditrapanij | |
| Nov 7, 2021 at 10:37 | vote | accept | A.A | ||
| Nov 7, 2021 at 4:34 | comment | added | ditrapanij | +1 to routing where possible. Also need to take into account that some providers MAY specifically block specific protocol control frames (e.g. STP BPDUs or even LLDP) through pseudowires. Best to ask your provider for technical specifications that details specific protocols and ether types that will be forwarded. | |
| Nov 6, 2021 at 10:47 | comment | added | Ricky | This is definitely a "your mileage my vary" situation. | |
| Nov 6, 2021 at 8:21 | history | answered | Zac67♦ | CC BY-SA 4.0 |