Package com.eqixiac.equinix.design.value


package com.eqixiac.equinix.design.value
The value-realization layer of the design API: design-time cost modelling that puts a money figure on an interconnect decision. It comprises the rate-card abstraction (com.eqixiac.equinix.design.value.ratecard), live cloud-provider pricing adapters (com.eqixiac.equinix.design.value.ratecard.provider), the egress savings calculator (com.eqixiac.equinix.design.value.savings), and the TCO comparison (com.eqixiac.equinix.design.value.tco).

Prices flow through a layered rate-card architecture: caller-supplied rates (CustomRateCard) take precedence over live Equinix Fabric pricing (EquinixRateCard), which takes precedence over bundled, dated reference figures (ReferenceRateCard); optional provider-API adapters slot in for live cloud egress rates. Every figure is tagged with its PriceSource so provenance survives aggregation, and a card that cannot price an item returns empty - never a phantom zero - so the next layer gets consulted.

The layer's one hard rule is currency honesty, centralized in CurrencyReconciler: amounts are only added, subtracted, or compared when they share a currency, and no FX rate is ever fabricated. When components disagree, the engines report the figure as unpriced or partial with per-currency subtotals and a reason, rather than emitting a false single-currency number. Entry points: fabric.savingsCalculator() and fabric.tcoComparison() (also available on Design.over(fabric)).

Cloud-to-cloud comparisons

Beta. Both entry points accept a peer cloud (toCloud(...)). The models then price the traffic between two clouds over three paths: the public internet, Equinix Fabric (private egress both ways, two virtual connections, both CSP ports), and the providers' native multicloud link, which has no Equinix component. The native link is priced by RateCard.multicloudLink(...) as a two-sided MulticloudLinkQuote: one flat fee per provider, converted from the published hourly rate at MulticloudLinkQuote.HOURS_PER_MONTH = 730 h/month.

Rules the cloud-to-cloud models follow
RuleMechanism
Every path is priced on the same two-way traffic and a complete component list a peer cloud switches every archetype to its two-sided form together; on-prem, whose inputs carry no cloud egress, leaves the default set and is reported unpriced when requested; a missing peer-side component marks the archetype partially priced
Only published native-link prices are bundledeach bundled figure records its source URL and retrieval date, repeated in the quote note; a size, tier, region or provider without a published figure is an empty side with a reason
The AWS free tier is never assumedapplied only on useAwsFreeTier(true)
The AWS path tier is never derivedan input, default 1; AWS publishes no path-to-tier table
No cross-currency sumsthe two sides of a link and the components of each path run through CurrencyReconciler

The savings estimate also reports the sustained rate at which the flat-fee native link and the per-GB Equinix path cost the same (MulticloudPathComparison.breakEvenSustainedMbps(...) documents the formula). Without a peer cloud, both models behave and render exactly as before.

See Also:
  • Class
    Description
    The single reconciliation rule shared by every design-time cost model (TcoEngine, SavingsCalculatorEngine, MetroOptimizerEngine, DeploymentWizardEngine, and plan value realization).
    The inputs the TCO and savings models take, derived from a DesignSpec (and optionally the portfolio a run chose) the same way the optimiser's RateCardCostModel reads them, so the three price the same thing.