Hello!
I'm struggling to understand the way Meraki does QoS/Traffic Shaping and am looking for some recommendations.
What I'm trying to understand is with an entire Meraki setup(Switches, APs, Security Appliance), with multiple locations(~25 users and ~10 phones per location) and dual wans(<10mbFiber/ >10mb "Business" class internet) at every location. What is the best way to setup "all low latency traffic and internal traffic exit wan fiber, all other data exit wan business. Would the traffic shaping rules only go on the firewall or would they go on the firewall and the access points? Would you implement bandwidth limits only on the firewall or on the firewall AND the APs?
Or is it better to setup everything to go out one WAN with rules to ensure VoIP traffic always prioritized and if failover needs to happen VoIP is still prioritized. I apologize for all the questions.
Thank you!
Traffic shaping can be used on Cisco Meraki MR and MX/Z devices to limit client throughput based on different rules. This allows less desirable traffic to be throttled while allowing more important traffic to be processed normally. It can also be used to limit the throughput for individual devices, to prevent a small number of clients from saturating the network, while other clients are unable to function.
https://documentation.meraki.com/MR/Firewall_and_Traffic_Shaping/Traffic_and_Bandwidth_Shaping
Okay awesome, would you recommend setting up the limits on both the firewall and access points or just the one? Are bandwidth limits better to use than just setting up rules with the traffic? Say VoIP or the VoIP VLAN is unlimited, but all youtube is 1Mb? Also, how does traffic shaping take into effect a dual wan setup? If you set each client to 10Mb because one of your WANs is 100Mb but the other is 5 and the line fails over how do you prevent issues? I apologize for so many questions, just in super unfamiliar territory. And how does this all get affected by the IPsec/VPN setup? My assumption is that if you setup VoIP as expedited traffic on one end, you'll 100% want to make it expedited on the other.
You can configure it just on MX. For another question I recommend reading this article:
Thank you for that, trying to read as much as possible and understand. We're not currently using load balancing, but set to failover with WAN preference of web traffic HTTP(S) over the high speed business class line, and the voice traffic over the lower speed fiber line. And I apologize, so if you have a line that's 150 down and 20 up, and then a 10/10 line and have global bandwidth limits set to 10 per client, how do you prevent one client from using an entire line, or is this where it's better to use traffic shaping instead?
The best option in my opinion is to set a bandwidth limit and then create rules prioritizing the most important applications. But note that there is no watering to be followed, it depends on your type of business and what your needs are.
Thank you for the information. My biggest concern is how does Meraki follow bandwidth limits on uneven line speeds. If I set a bandwidth limit for 10Mb and the 150Mb line goes down, I've now guaranteed that a single client can utilize the entire line even with prioritization if the rule is set to "obey the network limit per client" and the limit is 10Mb what's stopping anyone from utilizing the entire line themselves?