dispatcher: algorithm 13 latency optimized round-robin with failover

[jchavanton]

Type Of Change

• New feature (non-breaking change which adds new functionality)

Checklist:

• Tested changes locally

Description

Previously discussed in the mailing list :

“13” - latency optimized dispatching .

• The algorithm will load balance using round-robin prioritizing the gateways with the highest priority.
• If ds_ping_latency_stats is active the algorithm will adjust the priority of the gateway automatically,
  the priority will be lowered by 1 point every time the latency ms is as high as the priority.
• If the attribute 'cc=1' is set, the latency used is congestion ms : estimate (current latency ms) - average (normal condition latency ms).

Using this simple formula :
ADJUSTED_PRIORITY = PRIORITY - (ESTIMATED_LATENCY/PRIORITY)

GATEWAY	PRIORITY	ESTIMATED_LATENCY	ADJUSTED_PRIORITY	LOAD DISTRIBUTION
1	30	21	30	33%
2	30	91	27	0%
3	30	61	28	0%
4	30	19	30	33%
5	30	32	29	0%
6	30	0	30	33%
7	30	201	24	0%

With congestion control the formula becomes :

CONGESTION_MS = CURRENT_LATENCY_MS - NORMAL_CONDITION_LATENCY_MS
ADJUSTED_PRIORITY = PRIORITY - (CONGESTION_MS/PRIORITY)

[jchavanton]

• I will make sure the GW are well sorted, not only the first pick.
• I am considering additional support for weight based distribution with optional congestion control at equal priority.

[lgtm-com]

This pull request introduces 1 alert when merging f7c98e2 into 9935dc8 - view on LGTM.com

new alerts:

• 1 for Comparison result is always the same

[jchavanton]

This pull request introduces 1 alert when merging f7c98e2 into 9935dc8 - view on LGTM.com

new alerts:

* 1 for Comparison result is always the same

looked at it, hash is returned and can be -1 , unless I am blind

[lgtm-com]

This pull request introduces 1 alert when merging dbfe2d8 into e3e6471 - view on LGTM.com

new alerts:

• 1 for Comparison result is always the same

[lgtm-com]

This pull request introduces 1 alert when merging 46745e5 into a1620fa - view on LGTM.com

new alerts:

• 1 for Comparison result is always the same

[lgtm-com]

This pull request introduces 1 alert when merging b793b96 into e9624bc - view on LGTM.com

new alerts:

• 1 for Comparison result is always the same

[jchavanton]

Load tested Friday the 13th :)
250CPS

3 gateways, one with excessive ( latency >30ms delta with the 2 others since a priority of 30 is set)

URI: sip:65.75.65.247:5072
FLAGS: AP
PRIORITY: 30
ATTRS: {
	BODY: latency=10
}
LATENCY: {
	AVG: 10.039000
	STD: 1.381000
	EST: 10.001000
	MAX: 10
	TIMEOUT: 0
}

URI: sip:65.75.65.247:5071
FLAGS: AP
PRIORITY: 30
ATTRS: {
	BODY: latency=250
}
LATENCY: {
	AVG: 249.847000
	STD: 5.377000
	EST: 141.598000
	MAX: 250
	TIMEOUT: 0
}

URI: sip:65.75.65.247:5073
FLAGS: AP
PRIORITY: 30
LATENCY: {
	AVG: 0.000000
	STD: 0.000000
	EST: 0.000000
	MAX: 1
	TIMEOUT: 0
}

algorithm is doing round-robin on GW5072 and GW5073 and keep GW5071 for failover

 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[0]uri[sip:65.75.65.247:5072]priority[30-0=30]latency[10ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[1]uri[sip:65.75.65.247:5071]priority[30-8=22]latency[250ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[2]uri[sip:65.75.65.247:5073]priority[30-0=30]latency[0ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2224]: ds_manage_route_algo13(): priority[30]gateway_selected[0]next_index[1]

 1(50) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[1]uri[sip:65.75.65.247:5071]priority[30-8=22]latency[250ms]flag[8]
 1(50) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[2]uri[sip:65.75.65.247:5073]priority[30-0=30]latency[0ms]flag[8]
 1(50) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[0]uri[sip:65.75.65.247:5072]priority[30-0=30]latency[10ms]flag[8]
 1(50) INFO: dispatcher [dispatch.c:2224]: ds_manage_route_algo13(): priority[30]gateway_selected[2]next_index[0]

 5(54) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[0]uri[sip:65.75.65.247:5072]priority[30-0=30]latency[10ms]flag[8]
 5(54) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[1]uri[sip:65.75.65.247:5071]priority[30-8=22]latency[250ms]flag[8]
 5(54) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[2]uri[sip:65.75.65.247:5073]priority[30-0=30]latency[0ms]flag[8]
 5(54) INFO: dispatcher [dispatch.c:2224]: ds_manage_route_algo13(): priority[30]gateway_selected[0]next_index[1]

 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[1]uri[sip:65.75.65.247:5071]priority[30-8=22]latency[250ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[2]uri[sip:65.75.65.247:5073]priority[30-0=30]latency[0ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2207]: ds_manage_route_algo13(): [active]idx[0]uri[sip:65.75.65.247:5072]priority[30-0=30]latency[10ms]flag[8]
 2(51) INFO: dispatcher [dispatch.c:2224]: ds_manage_route_algo13(): priority[30]gateway_selected[2]next_index[0]

Total 20000 INVITE calls sent in 79433 ms at rate of 251/sec
Total 20000 connection responses received in 82101 ms at rate of 243/sec:
>> Detailed connection responses received:
 - 200 connection responses:    20000     (OK)
>> Detailed disconnection responses received:
 - 200 disconnection responses:    20000     (OK)
                    ------
 TOTAL responses:    20000 (rate=243/sec)

Maximum outstanding job: 719

[henningw]

Thank you for the pull request, and also the announcement on the list.
Are there any more additions/refinements done from your side? Otherwise if there are no objections it can be merged, i think.