Hi
I’m new to MikroTik routers, just purchased, using Winbox running OS v7.1.2 on RB5009UG+S+(arm64)
I’m been reading the Simple Queue Manual, some topics on forum and watching several youtube videos on the subject, but seems a bit mixed / conflicting info.
many of the guides and videos nicely demonstrates Simple Queue, but seems to me this limiting is a bit unclear to WAN vs. LAN and as well committing vs
max, and if max is the total for all or as for each client when not targeting each client separately.
Setup
ISP, dynamic, fiber 1000 down (reality around 980) / 500 up (reality about 550) on ether 1 (2,5 GB port)
bridge1 internal - ether 2,3,4 (1 GB ports) - DHCP server 192.168.6.0/24
(PC’s, lap tops, file-server, printer, via some unmanaged Cisco switches)
bridge2 guests - ether 5,6,7 (1 GB ports) - DHCP server 192.168.7.0/24
(smartTV and AP’s, using some Asus routers as AccessPoint’s wireless and switch)
bridge3 game - ether 8 - (1 GB port) - DHCP server 192.168.8.0/24
(PC’s and unmanaged switch)
expect a max of approx 30 concurrent users
I’m trying to avoid VLAN/packet marking as i’m not sure packet markings will effectively pass trough all the unmanaged switches I have
Concern:
WAN speed is close to saturating limit of general network ethernet 1GB ports,
i.e. internal traffic peer to peer, file server, printer will as well load the network and may limit/delay internet.
Aim:
To keep rules/queue as simple as possible for traffic limiting/shaping, and not to micromanage each client IP separately (except maybe file server and printer)
To utilize the router / multi core processor best possible for bandwidth limit/shaping, i.e. simple queue / parent set up enabling multi core usage.
To as far as possible use the OS Windbox (or Webfig) as the tool, i.e. avoid terminal scripts
Goal:
WAN
- to cap that any client can exceed a max of 500 down / 400 up
(i.e. if a user runs a bittorrent, speed test, or fast cloud/ftp up/down, they will initially never load the internet to the actual available limit) - to shape any client to guarantee a priority of minimum of 30 down / 15 up, but otherwise if bandwidth is available utilize up the max 500/400
(i.e. any client will always obtain possible lowest latency to/from internet up to the minimum guaranteed speed before buffering/delays) - to priority in order game bridge clients - internal bridge clients - guests bridge clients
(i.e. should congestion / buffering appear the priority to lowest latency is in order game bridge users first, internal next and then guest)
LAN - to cap that any client can exceed a max of 500 down / 500 up on the switching / LAN
- to shape equally client on internal LAN if more users are utilizing this
- to priorities internal networking on LAN as 2nd priority to WAN
What’s done, and questions I like to get a little guidance on or confirmed
Setting a parent “WANall” for overall WAN Target=empty(all) / Dest=ethernet1 Max down 900 / Max up 500, Advanced Priority 1 (telling the router the available speed for shaping)
Setting a child "WANgame) parent=WANall, Target=bridge3 / Dest=ethernet1 Max down 500 / Max up 400, Advanced no packet marking Limit at up=30 / down=15 Priority 2
Setting a child "WANinternal) parent=WANall, Target=bridge1 / Dest=ethernet1 Max down 500 / Max up 400, Advanced no packet marking Limit at up=30 / down=15 Priority 3
Setting a child "WANguest) parent=WANall, Target=bridge2 / Dest=ethernet1 Max down 500 / Max up 400, Advanced no packet marking Limit at up=30 / down=15 Priority 4
Setting a parent “LANall” for overall LAN Target=brige1, bridge2,bridge3 / Dest=empty(all) Max down 1000 / Max up 1000 Priority 5 (telling the router the available speed for shaping)
Setting a child "LANgame) parent=LANall, Target=bridge3 / Dest=empty(all) Max down 500 / Max up 500, Advanced no packet marking Limit at up=30 / down=15 Priority 6
Setting a child "LANinternal) parent=LANall, Target=bridge1 / Dest=empty(all) Max down 500 / Max up 500, Advanced no packet marking Limit at up=30 / down=15 Priority 7
Setting a child "LANguest) parent=LANall, Target=bridge2 / Dest=empty(all) Max down 500 / Max up 500, Advanced no packet marking Limit at up=30 / down=15 Priority 8
Disabling the default fast-track firewall rule, to enable the limits and shaping, is this correct ?
Will this obtain the goal of single clients able to utilizing available bandwidth up to max setting and as well guarantee each user a minimum speed/low latency/no-buffering when network is loaded ?
Do i need to address a simple rule for the AP’s separately, as the AP will actually have an IP assigned as well, if considered as a client in the que, would this limit to a total of clients connected through the AP, i.e. AP own traffic, backhaul etc. if the AP as a client “only” has a 30/15 guaranteed to serve all the clients connected to the AP ?
Can I use bridge as defined targets and destinations, (will it apply to each client separately or is it a limit to the entire bridge as a total), OR do i need list the ports (maybe same issue as bridge)
Or is it needed to set the IP range instead to make it apply to each client in the IP range ?
Will this actually prioritize the WAN above LAN ?
Will a VPN form a PC to a cooperate network outside be catch in the que rules ?
Do I have with the setup queues having risk of double traffic limits (performing limits twice) ?
Finally, the WAN from ISP is currently “normal” way of being assigned public IP and DNS, but I’m about to change ISP, and this will require to “search” VLAN10 to obtain the IP and DNS set, i.e. some implementation of VLAN, will this change anything of above, as by now I run the separation solely by separate DHCP servers on subnets.