Routing is the process of selecting paths across the networks to move packets from one host to another.
Let's look at a basic configuration example to illustrate how routing is used to forward packets between two local networks and to the Internet.
In this setup, we have several networks:
Router1 (gateway) where ether1 connects to the internet:
If we look, for example, at the Router1 routing table, we can see that the router knows only about directly connected networks. At this point, when Client from LAN1 tries to reach client from LAN2 (192.168.2.0/24), a packet will be dropped on the router, because destination is unknown for the particular router:
To fix this we need to add a route which tells the router what is the next device in the network to reach the destination. In our example next hop is Router2, so we need to add a route which gateway will point to the Router's 2 connected address:
At this point packet from LAN1 will be successfully forwarded to LAN2, but we are not over yet. Router2 does not know how to reach LAN1, so any packet from LAN2 will be dropped on Router2.
If we look again at the network diagram, we can clearly see that Router2 has only one point of exit. It is safe to assume that all other unknown networks should be reached over the link to Router1. The easiest way to do this is by adding a default route: To add default route specify destination 0.0.0.0/0 or leave it blank:
As we have seen from example setup, there are different groups of routes, based on their origin and properties.
RouterOS routing information consists of two main parts:
Routing Information Base is a database that lists entries for particular network destinations and their gateways (address of the next device along the path or simply nexthop). One such entry in the routing table is called route.
A hop occurs when a packet is passed from one network segment to another.
By default, all routes are organized in one "main" routing table. It is possible to make more than one routing table which we will discuss further in this article, but for now, for sake of simplicity, we will consider that there is only one "main" routing table.
RIB table contains complete routing information, including static routes and policy routing rules configured by the user, routing information learned from dynamic routing protocols (RIP, OSPF, BGP) and information about connected networks.
Its purpose is not just to store routes, but also to filter routing information to calculate the best route for each destination prefix, to build and update Forwarding Information Base and to distribute routes between different routing protocols.
Connected routes represent the network on which hosts can be directly reached (direct attachment to Layer2 broadcast domain). These routes are created automatically for each IP network that has at least one enabled interface attached to it (as specified in the /ip address or /ipv6 address configuration). RIB tracks the status of connected routes but does not modify them. For each connected route there is one IP address item such that:
The preferred source is not used anymore for connected routes. FIB chooses source address based on the out-interface. This allows making setups that in ROS v6 and older were considered invalid. See examples for more details.
A default route is used when the destination cannot be resolved by any other route in the routing table. In RouterOS dst-address of default route is 0.0.0.0/0 (for IPv4) and ::/0 (for IPv6) routes. If the routing table contains an active default route, then the routing table lookup in this table will never fail.
Typically home router routing table contains only connected networks and one default route to forward all outgoing traffic to ISPs gateway:
To implement some setups, such as load balancing, it might be necessary to use more than one path to a given destination.
ECMP (Equal cost multi-path) routes have multiple gateways (nexthop) values. All reachable nexthops are copied to FIB and are used to forward packets.
These routes can be created manually, as well as dynamically by any of the dynamic routing protocols (OSPF, BGP, RIP). Multiple equally preferred routes to the same destination will have assigned + flag and grouped together automatically by RouterOS (see example below).
There can be multiple routes with the same destination received from various routing protocols and from static configuration but only one (best) destination can be used for packet forwarding. To determine the best path, RIB runs Route Selection algorithm which picks best route from all candidate routes per destination.
Only routes that meet following criteria can participate in route selection process:
Candidate route with the lowest distance becomes an active route. If there is more than one candidate route with the same distance, selection of active route is arbitrary.
Nexthop lookup is a part of the route selection process. Its main purpose is to find directly reachable gateway address (nexthop). Only after valid nexthop is selected router knows which interface to use for packet forwarding.
Nexthop lookup becomes more complicated if routes have gateway address that is several hops away from this router (e.g. iBGP, multihop eBGP). Such routes are installed in the FIB after nexthop selection algorithm determines the address of directly reachable gateway (immediate nexthop).
It is necessary to restrict set of routes that can be used to look up immediate nexthops. Nexthop values of RIP or OSPF routes, for example, are supposed to be directly reachable and should be looked up only using connected routes. This is achieved using scope and target-scope properties.
Routes with scope greater than the maximum accepted value are not used for nexthop lookup. Each route specifies maximum accepted scope value for it's nexthops in the target-scope property. Default value of this property allows nexthop lookup only through connected routes, with the exception of iBGP routes that have larger default value and can lookup nexthop also through IGP and static routes.
There are changes in RouterOS v7 nexthop lookup.
Routes are processed in scope order, and updates to routes with larger scope cannot affect state of nexthop lookup for routes with smaller scope.
Consider an example from v6:
Gateway 10.0.0.1 is recursively resolved through C using smallest referring scope (scope 20 from route B), both routes are active. Now we change both A and B at the same time:
Suddenly, applying update of route A makes gateway of route B inactive. This is because in v6 there is only one gateway object per address.
v7 keeps multiple gateway objects per address, one for each combination of scope and gateway-check.
Changing target-scope or gateway-check of a route in v7 will not affect other routes, as it does in v6. In v7 target-scope and gateway-check are properties that are internally attached to the gateway, not to the route.
Routing information is stored to take as little memory as possible in a common case. These optimizations have non-obvious worst-cases and impact on perfomance.
All routes and gateways are kept in a single hierarchy by they prefix/address.
Dst /0 1/0+4 18 <-- number of prefixes ^ ^ ^ ^ ^ | | | | | | | | | \- bytes taken by Route distinguisher or Interface Id | | | \--- vrf/routing table | | \----- AFI | \------- netmask length of prefix \---------- bytes taken by prefix value [stuff subject to change without notice]
Each of these 'Dst' corresponds to a unique 'dst-address' of route or address of gateway. Each 'Dst' requires one or more 'T2Node' objects as well.
All routes with the same 'dst-address' are kept in Dst in a list sorted by route preference.
Note: WORST CASE: having a lot of routes with the same 'dst-address' is really slow! even if they are inactive! because updating sorted list with tens of thousands of elements is slow!
Route order changes only when route attributes change. If route becomes active/inactive, order does not change.
Each Route has three copies of route attributes:
!!!!! Need option to print private !!!!!
Periodically (when needed), update attributes are calculated from private attributes. This happens when route update is received, or when in-filter is updated.
When routing table is recalculated, current attributes are set to the value from updated attributes.
This means, that usually, if there is no in-filter that changes route attributes, private, updated and current share the same value.
Route attributes are kept in several groups:
Having for example many different combinations of distance and scope route attributes will use more memory!
Matching communities or as-path using regexp will cache the result, to speed up filtering. Each as-path or community value has cache for all regexps, which is filled on demand with match results.
Note: WORST CASE: changing attributes in 'in-filter' will make route program use more memory! Because 'private' and 'updated' attributes will be different! Having a lot of different regexps will make matching slow and use a lot of memory! Because each value will have a cache with thousands of entries!
Detailed info about used memory by routing protocols can be seen in
/routing stats memory menu
FIB (Forwarding Information Base) contains copy of information that is necessary for packet forwarding:
Each route has dst-address property, that specifies all destination addresses this route can be used for. If there are several routes that apply to a particular IP address, the most specific one (with largest netmask) is used. This operation (finding the most specific route that matches given address) is called ''routing table lookup''.
Only one Best route can be used for packet forwarding. In cases where routing table contains several routes with the same dst-address, all equally best routes are combined into one ECMP route. Best route is installed into FIB and marked as ''active''.
When forwarding decision uses additional information, such as a source address of the packet, it is called policy routing. Policy routing is implemented as a list of policy routing rules, that select different routing table based on destination address, source address, source interface, and routing mark (can be changed by firewall mangle rules) of the packet.
FIB uses following information from packet to determine it's destination:
Possible routing decisions are:
Run routing decision:
Result of routing decision can be:
Rules that do not match current packet are ignored. If rule has action:
In RouterOS you have three menus to see current state of routes in routing table:
/ip route- list IPv4 routes and basic properties
/ipv6 route- list IPv6 routes and basic properties
/routing route- list all routes with extended properties
/routing route menu currently is read only. To add or remove routes
/ip(ipv6) route menus should be used.
[admin@MikroTik] /ip/route> print Flags: D - dynamic; X - disabled, I - inactive, A - active; C - connect, S - stati
c, r - rip, b - bgp, o - ospf, d - dhcp, v - vpn
Columns: DST-ADDRESS, GATEWAY, DIstance
# DST-ADDRESS GATEWAY DI
0 XS 10.155.101.0/24 188.8.131.52
1 XS 184.108.40.206
D d 0.0.0.0/0 10.155.101.1 10
2 AS 0.0.0.0/0 10.155.101.1 1
3 AS 220.127.116.11/24 10.155.101.1 10
4 AS 18.104.22.168 22.214.171.124 1
DAC 10.155.101.0/24 ether12 0
| ||| | | | | ||| | | \----Distance | ||| | \--Configured gateway | ||| \-- dst prefix | ||\------- protocol flag (bgp, osf,static,connected etc.) | |\-------- route status flag (active, inactive, disabled) | \--------- shows if route is dynamic \----------- console order number (shown only for static editable routes)
routing route output is very similar to ip route except that it shows routes from all address families in one menu and list filtered routes as well.
[admin@MikroTik] /routing/route> print Flags: X - disabled, I - inactive, F - filtered, U - unreachable, A - active; c - connect, s - static,
r - rip, b - bgp, o - ospf, d - dhcp, v - vpn, a - ldp-address, l - ldp-mapping
Columns: DST-ADDRESS, GATEWAY, DIStance, SCOpe, TARget-scope, IMMEDIATE-GW
DST-ADDRESS GATEWAY DIS SCO TAR IMMEDIATE-GW
d 0.0.0.0/0 10.155.101.1 10 30 10 10.155.101.1%ether12
As 0.0.0.0/0 10.155.101.1 1 30 10 10.155.101.1%ether12
As 126.96.36.199/24 10.155.101.1 10 30 10 10.155.101.1%ether12
As 188.8.131.52 184.108.40.206 1 254 254 10.155.101.1%ether12
Ac 10.155.101.0/24 ether12 0 10 ether12
Ic 2001:db8:2::/64 ether2 0 10
Io 2001:db8:3::/64 ether12 110 20 10
Ic fe80::%ether2/64 ether2 0 10
Ac fe80::%ether12/64 ether12 0 10 ether12
Ac fe80::%bridge-main/64 bridge-main 0 10 bridge-main
A ether12 0 250
A bridge-main 0 250
routing route print detail shows more advanced info useful for debugging
[admin@MikroTik] /routing route> print detail
Flags: X - disabled, I - inactive, F - filtered, U - unreachable, A - active;
c - connect, s - static, r - rip, b - bgp, o - ospf, d - dhcp, v - vpn, a - ldp-address, l - ldp-ma>
+ - ecmp
d afi=ip4 contribution=best-candidate dst-address=0.0.0.0/0 gateway=10.155.101.1
immediate-gw=10.155.101.1%ether12 distance=10 scope=30 target-scope=10
belongs-to="DHCP route" mpls.in-label=0 .out-label=0 debug.fwp-ptr=0x201C2000
As afi=ip4 contribution=active dst-address=0.0.0.0/0 gateway=10.155.101.1
immediate-gw=10.155.101.1%ether12 distance=1 scope=30 target-scope=10
belongs-to="Static route" mpls.in-label=0 .out-label=0 debug.fwp-ptr=0x201C2000