Struct QueryCost 

pub struct QueryCost {
    pub traversal_s: f64,
    pub storage_access_s: f64,
    pub distance_s: f64,
}

The three-term query-cost decomposition, in seconds.

Produced by QueryCost::estimate, which prices a set of QueryCounters against a StorageProfile and a DistanceCost. The point of keeping the terms separate (rather than only their sum) is that the storage-aware claims are about how a change shifts cost between terms — e.g. a narrower dtype moves cost out of storage_access_s, a larger approximate gate moves cost out of storage_access_s and into distance_s.

Fields

traversal_s: f64

c_hop · |V| — best-first bookkeeping over visited nodes (RAM-bound).

storage_access_s: f64

Σ (t_seek + bytes/B_seq) — physical reads. The term prior work assumes uniform; the dominant term on HDD-class media.

distance_s: f64

c_dist · (|A| + |E|) — approximate plus exact distance computation.

Implementations

impl QueryCost

pub const HOP_OVERHEAD_S: f64 = 50e-9

Per-visited-node traversal overhead, in seconds.

Best-first bookkeeping (queue ops, adjacency expansion, visited-set maintenance) is small and RAM-bound; this is a representative constant, not a tuned value. Traversal is intentionally the cheap term in the storage-aware regime.

pub fn estimate( counters: &QueryCounters, profile: StorageProfile, cost: DistanceCost, ) -> Self

Price counters against a storage profile and distance cost.

Storage access is modelled as one seek per seeks counted plus transfer of bytes_read at the medium’s sequential bandwidth — so a query that coalesces its reads into fewer seeks (via layout + scheduling) is priced strictly lower at equal bytes, which is the behaviour the paradigm predicts and the scheduler is built to produce.

pub fn total_s(self) -> f64

Total modelled query latency: the sum of the three terms.

pub fn storage_fraction(self) -> f64

Fraction of total cost attributable to storage access, in [0, 1].

A diagnostic for which regime a query is in: near 1 means the query is storage-bound (the paradigm’s target regime), near 0 means it is compute- or traversal-bound (the classical in-RAM regime).

Trait Implementations

impl Clone for QueryCost

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for QueryCost

impl Debug for QueryCost

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for QueryCost

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for QueryCost

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for QueryCost

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for QueryCost