I would put it differently (but to the same effect ).
At the time Mikrotik devices mostly had 10/100 ethernet ports, having the CHR limited to 1 Mbps meant having it 1/100 of max speed available on most hardware.
Now that Mikrotik devices mostly have at least 1 Gbps ethernet ports making the CHR limited to 10 Mbps would mantain the tradition of 1/100.
Totally in agreement
With the amount of testing i have todo its become incredible painful to try test, benchmarks, system resource requirements, Bandwidth QOS , Que trees etc etc on 1M
I would be ok with a "Technician License" at a cheap rate that would require regular relicensing. The Trial doesnt cut it when you having to test multiple device setups, over several months
From mikrotik's side the license can literally stop working after x period and require a renewal/rest period to avoid people trying to use this in production
On a side funny note. I got asked by CFO if we could get Licenses so we could virtualizes all the Wifi testing .
If the free license offered 10 Mbps, I'd have not two licensed P1 CHRs, but probably 20 unlicensed CHRs. My average traffic is 5-7 Mbps. 10 Mbps would turn CHRs into a free (3$) personal VPN, for those who desperately need a VPN but are prevented from purchasing a license by sanctions.
Buy less coffee and make it yourself. Return some bottles, the cost of lifetime license is not that much.
If tis really important then you can make it happen.
There already is the P1 license that offers 1Gbps and costs like 35 euro when you shop around. Maybe more reasonable would be to ask for a “10-pack P1 license” (or maybe 5-pack P1 license) at an interesting discount.
(in the past the interesting alternative was to visit a MUM. one day of interesting talks, free lunch and drinks, merch, free device (usually some wAP or lite router), and a P1 license, all for 40 euro)
Agreed; I’d suggest alternatives of 3 or 5Mbps, or at least a tweak to the shaper to allow for short bursts of traffic (5-10 seconds) commonly seen during BGP convergence.
Another idea might be a shortened version of the 24-hr window offered for hardware-based installations. Long enough to support spinning up VMs, loading configs, testing the scenario, then shutting down the lab (6-12 hours?).
Or an out-of-band/romon-style lab licensing server/proxy with which all VM’s check in. A user buys 10-20 lab licenses, and the server leases those licenses like a DHCP lease.
Part of this issue is that it's actually not that easy to correctly automate the licensing, the call is asynchronous (returns status pages, you see the paging in CLI, but really problematic from REST API)... so to KNOW you licensed it and check for errors is bit trickier than it should be too. I've started licensing with trial some of my tikoci GitHub project, since it does speed up things like crawling APIs (100K+ http calls, going via 1Mb/s, works but takes 3-5x as long) – so familiar with this problem.
And, there is up limit to 100 trial license. I know since my account did hit the limit, since I was testing a script to do the license.... Really tricky bug since licensing from REST returned a 200. But my script read back the status, and it was still unlicensed. So you have to parse the JSON (trolling by doing string matching & looking for last result) returned to know if it was actually successful.
I'm working a tool called quickchr that automated bring up a provisioned CHR with licensing, device-mode, admin user, and packages all set & wired up to networks, using pure QEMU. I'll post that soon. It has both a wizard to bring up a CHR - including licensing - as well as CLI (and TypeScript API). Basically Ubuntu's multipass for CHR. I'll post it soon.
TEASER
And yes it handles the failure mode of >100 licenses:
You missed the point. I wasn't writing about the cost of CHR, I was writing about the cost of a VPS for CHR that doesn't require a license. The point was that sanctions prevent people from purchasing a CHR license. And if the speed will 10 Mbps without a license, people will happily use such a CHR.
).
