Struct Timers 

pub struct Timers {
    pub ta: Duration,
    pub pacing_floor: Duration,
    pub rto_floor: Duration,
    pub rc: u32,
    pub rm: u32,
    pub tr: Duration,
    pub tn: Duration,
}

The timers of RFC 8445 §14 and RFC 5389 §7.2.1, plus the one stopping value that is sipx’s own.

Every field is a value the deployment may change; Timers::default is what the RFCs recommend. Two of them have normative floors that Timers::pacing and Timers::rto enforce rather than trust: Ta may not pace faster than 5 ms across every agent in the process (§14.2), and an RTO may never be below 500 ms (§14.3).

Fields

ta: Duration

Ta — the pacing interval: one check leaves per tick, across the whole checklist set (§14.2). Default 50 ms.

pacing_floor: Duration

The floor under Ta, “as though there were one global Ta value for pacing all agents” (§14.2). Default 5 ms.

rto_floor: Duration

The floor under the RTO. §14.3: agents “MUST NOT use an RTO value smaller than 500 ms”.

rc: u32

Rc — how many times a request is transmitted before the transaction fails (RFC 5389 §7.2.1). Default 7.

rm: u32

Rm — the multiplier on the wait after the last transmission (RFC 5389 §7.2.1). Default 16.

tr: Duration

Tr — how long a selected pair may carry no data before a keepalive is sent (§11). Default 15 s, which §11 also makes the minimum: “MUST NOT use a value smaller”.

tn: Duration

Tn — sipx’s own stopping value: how long the controlling agent keeps checking after the first valid pair appears before it nominates (spec §8). Default 1 s.

§8.1.1 leaves the stopping criterion to local optimisation and requires only that exactly one pair is eventually nominated. A number here rather than an emergent behaviour is what makes that choice testable.

Implementations

impl Timers

pub fn pacing(&self) -> Duration

The interval between checks: Ta, but never below the process-wide floor (§14.2).

pub fn rto(&self, checks: usize, outstanding: usize) -> Duration

The retransmission interval for a check being sent now (§14.3):

RTO = MAX(500ms, Ta * N * (Num-Waiting + Num-In-Progress))

checks is N, the total number of connectivity checks to be performed — the size of the checklist set, not of one checklist. outstanding is Num-Waiting + Num-In-Progress across that same set, which is why this cannot be computed once: it falls as the checks drain, and §14.3 says so outright.

pub fn final_wait(&self, rto: Duration) -> Duration

The wait after the last transmission of a check, after which the transaction has timed out (RFC 5389 §7.2.1: “a duration equal to Rm times the RTO”).

pub fn double(&self, rto: Duration) -> Duration

The next retransmission interval: RFC 5389 §7.2.1 doubles it after every transmission.

Trait Implementations

impl Clone for Timers

fn clone(&self) -> Timers

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Timers

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Timers

fn default() -> Self

Returns the “default value” for a type.

impl PartialEq for Timers

fn eq(&self, other: &Timers) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for Timers

impl Eq for Timers

impl StructuralPartialEq for Timers