Lowest cost (SLA) strategy
When using Lowest Cost (SLA) mode (sla
in the CLI), SD-WAN will choose the lowest cost link that satisfies SLA to forward traffic. The lowest possible cost is 0
. If multiple eligible links have the same cost, the Interface preference order will be used to select a link.
In this example, your wan1 and wan2 SD-WAN interfaces connect to two ISPs that both go to the public internet. The cost of wan2 is less than that of wan1. You want to configure Gmail services to use the lowest cost interface, but the link quality must meet a standard of latency: 10ms, and jitter: 5ms.
To configure an SD-WAN rule to use Lowest Cost (SLA):
- On the FortiGate, add wan1 and wan2 as SD-WAN members, then add a policy and static route. See SD-WAN quick start for details.
- Go to Network > SD-WAN, select the Performance SLAs tab, and click Create New.
- Enter a name for the performance SLA, such as google, and set the Server to google.com.
- Enable SLA Target. Set the Latency threshold to 10 ms, and the Jitter threshold to 5 ms. See Health checks for more details.
- Click OK.
- Go to Network > SD-WAN, select the SD-WAN Rules tab, and click Create New.
- Enter a name for the rule, such as gmail.
- Configure the following settings:
Internet Service
Google-Gmail
Strategy
Lowest Cost (SLA)
Interface preference
wan1 and wan2
Required SLA target
google
- Click OK.
To configure an SD-WAN rule to use SLA:
config system sdwan config members edit 1 set interface "wan1" set cost 10 next edit 2 set interface "wan2" set cost 5 next end config health-check edit "google" set server "google.com" set members 1 2 config sla edit 1 set latency-threshold 10 set jitter-threshold 5 next end next end config service edit 1 set name "gmail" set mode sla set internet-service enable set internet-service-id 65646 config sla edit "google" set id 1 next end set priority-members 1 2 next end end
If no SD-WAN zone is specified, members are added to the default virtual-wan-link zone. |
The CLI command |
To diagnose the performance SLA status:
FGT # diagnose sys sdwan health-check status Health Check(google): Seq(1): state(alive), packet-loss(0.000%) latency(14.563), jitter(4.334) sla_map=0x0 Seq(2): state(alive), packet-loss(0.000%) latency(12.633), jitter(6.265) sla_map=0x0 FGT # diagnose sys sdwan service 1 Service(1): Address Mode(IPV4) flags=0x0 TOS(0x0/0x0), Protocol(0: 1->65535), Mode(sla) Members:<<BR>> 1: Seq_num(2), alive, sla(0x1), cfg_order(1), selected 2: Seq_num(1), alive, sla(0x1), cfg_order(0), selected Internet Service: Google.Gmail(65646)
When both wan1 and wan2 meet the SLA requirements, Gmail traffic will only use wan2. If only wan1 meets the SLA requirements, Gmail traffic will only use wan1, even though it has a higher cost. If neither interface meets the requirements, wan2 will be used.
If both interface had the same cost and both met the SLA requirements, the first link configured in set priority-members
would be used.
Load balancing strategy with SLA targets
SD-WAN rules can be configured to load balance traffic out of all the interfaces that satisfy the SLA target.
The load balancing strategy known as maximize bandwidth (load-balance
) prior FortiOS 7.4.1 is now configured within Lowest Cost (SLA) mode (sla
) SD-WAN rules.
By enabling load balancing mode (set load-balance enable
) inside the lowest cost SD-WAN rule, SD-WAN will choose all of the links that satisfy the SLA target to forward traffic based on a load balancing algorithm. The load balancing algorithm, or hash method, can be one of the following:
round-robin |
All traffic is distributed to selected interfaces in equal portions and circular order. This is the default method, and the only option available when using the GUI. |
source-ip-based |
All traffic from a source IP is sent to the same interface. |
source-dest-ip-based |
All traffic from a source IP to a destination IP is sent to the same interface. |
inbandwidth |
All traffic is distributed to a selected interface with most available bandwidth for incoming traffic. |
outbandwidth |
All traffic is distributed to a selected interface with most available bandwidth for outgoing traffic. |
bibandwidth |
All traffic is distributed to a selected interface with most available bandwidth for both incoming and outgoing traffic. |
When the inbandwidth
, outbandwidth
, or bibandwidth
load balancing algorithm is used, the FortiGate will compare the bandwidth based on the configured upstream and downstream bandwidth values.
The interface speedtest can be used to populate the bandwidth values based on the speedtest results. See GUI speed test for details.
To manually configure the upstream and downstream bandwidth values:
config system interface edit <interface> set estimated-upstream-bandwidth <speed in kbps> set estimated-downstream-bandwidth <speed in kbps> next end
Example
Based on the same topology as the preceding example, your wan1 and wan2 SD-WAN interfaces connect to two ISPs that both go to the public internet. You want to configure Gmail services to use both of the interface, but the link quality must meet a standard of latency: 10ms, and jitter: 5ms. This can maximize the bandwidth usage by using load balancing.
To configure an SD-WAN rule to use load balancing with SLA targets in the GUI:
- On the FortiGate, add wan1 and wan2 as SD-WAN members, then add a policy and static route. See SD-WAN quick start for details.
- Go to Network > SD-WAN, select the Performance SLAs tab, and click Create New.
- Enter a name for the performance SLA, such as google, and set the Server to google.com.
- Enable SLA Target. Set the Latency threshold to 10 ms, and the Jitter threshold to 5 ms. See Health checks for more details.
- Click OK.
- Go to Network > SD-WAN, select the SD-WAN Rules tab, and click Create New.
- Enter a name for the rule, such as gmail.
- Configure the following settings:
Internet Service
Google-Gmail
Strategy
Lowest Cost (SLA)
Interface preference
wan1 and wan2
Required SLA target
google
Load balancing
Enable this setting
- Click OK.
To configure an SD-WAN rule to use load balancing with SLA targets in the CLI:
config system sdwan config members edit 1 set interface "wan1" set cost 10 next edit 2 set interface "wan2" set cost 5 next end config health-check edit "google" set server "google.com" set members 1 2 config sla edit 1 set latency-threshold 10 set jitter-threshold 5 next end next end config service edit 1 set name "gmail" set load-balance enable set mode sla set internet-service enable set internet-service-name "Google-Gmail" config sla edit "google" set id 1 next end set priority-members 1 2 next end end
To diagnose the performance SLA status:
FGT # diagnose sys sdwan health-check status Health Check(google): Seq(1): state(alive), packet-loss(0.000%) latency(14.563), jitter(4.334) sla_map=0x0 Seq(2): state(alive), packet-loss(0.000%) latency(12.633), jitter(6.265) sla_map=0x0 FGT # diagnose sys sdwan service 1 Service(1): Address Mode(IPV4) flags=0x0 TOS(0x0/0x0), Protocol(0: 1->65535), Mode(load-balance) Members:<<BR>> 1: Seq_num(1), alive, sla(0x1), num of pass(1), selected 2: Seq_num(2), alive, sla(0x1), num of pass(1), selected Internet Service: Google.Gmail(65646)
When both wan1 and wan2 meet the SLA requirements, Gmail traffic will use both wan1 and wan2. If only one of the interfaces meets the SLA requirements, Gmail traffic will only use that interface.
If neither interface meets the requirements but the health-check
is still alive, then wan1 and wan2 tie. The traffic will try to balance between wan1 and wan2, using both interfaces to forward traffic.
The maximize bandwidth (
|