diff --git a/documentation/OED-predeployment-viability-enhancement.md b/documentation/OED-predeployment-viability-enhancement.md new file mode 100644 index 0000000..76d5d5b --- /dev/null +++ b/documentation/OED-predeployment-viability-enhancement.md @@ -0,0 +1,150 @@ +# API Scope Enhancement Template + +## API name + +Optimal Edge Discovery (OED) +[https://github.com/camaraproject/OptimalEdgeDiscovery](https://github.com/camaraproject/OptimalEdgeDiscovery) + +## New API name + +N/A (scope extension of OED) + +## Scope Enhancement owner + +University of Patras + +## Scope Enhancement summary + +This proposal explores extending Optimal Edge Discovery (OED) to support **pre-deployment viability discovery**, in addition to its current runtime optimization capabilities. + +Today, OED focuses on selecting the optimal edge zone for an active device or session context. This enhancement introduces support for **deployment planning queries**, where no active device is present and the goal is to determine which edge options are viable in a target area given an application profile and constraints. + +### Example use cases + +* A cloud gaming provider evaluates whether a service can be launched in a specific geographic area based on latency constraints. +* An application provider compares multiple regions/operators to identify viable rollout locations before deployment. +* An aggregator platform provides developers with a unified interface to assess edge availability across multiple operators. + +--- + +## Delimitation versus existing APIs + +| API | Decision supported | Timing | Input context | Output | +| ------------------------ | --------------------------------- | ----------------------- | ----------------------------------------------- | ------------------------------------------------- | +| **OED** | Best edge selection | Runtime / near-runtime | Active device / session | Single optimal edge zone | +| **AED** | Application connectivity | Runtime | Deployed application | Endpoint information | +| **EAM** | Deployment & lifecycle | Deployment / operations | Application package | Deployment actions | +| **Proposed enhancement** | Deployment feasibility (go/no-go) | Pre-deployment | Target area + application profile + constraints | Set of viable edge options with estimated metrics | + +**Key distinction:** +OED answers *“what is the best edge now?”* +This enhancement answers *“is deployment feasible here at all?”* + +--- + +## Technical viability + +The enhancement relies on capabilities already aligned with current CAMARA and 3GPP architectures: + +* Exposure of network performance characteristics (e.g. latency estimates) through operator APIs (e.g. NEF-based exposure) +* Edge/cloud infrastructure awareness (edge zones, providers) +* Application profile–based constraint evaluation (already part of OED context) + +No fundamentally new network capabilities are required. The proposal extends the discovery model from **runtime target selection** to **pre-deployment feasibility evaluation**, reusing: + +* performance estimation mechanisms already used for edge discovery +* application profile constraints already present in OED +* operator or platform knowledge of edge zone capabilities + +Standards alignment: + +* 3GPP NEF (Rel-15+) for network exposure +* ETSI MEC / edge platform exposure models (where applicable) +* CAMARA Commonalities for geographic/context modeling + +--- + +## Minimal API contract sketch + +### Operation + +`POST /edge-discovery/viability` + +### Inputs + +* `targetArea`: geographic area identifier (e.g. region, city) +* `applicationProfile`: application requirements or reference profile +* `constraints`: mandatory requirements (e.g. max latency p95) +* `preferences` (optional): preferred providers or deployment hints + +### Outputs + +* `edgeOptions`: list of candidate edge zones +* `viability`: whether the candidate satisfies all mandatory constraints +* `estimatedMetrics`: planning-oriented metrics (e.g. latency p95, latency range) +* `providerInfo`: developer-facing provider/operator identifiers + +### Meaning of “viable” + +A candidate is considered **viable** if all mandatory constraints defined in the request are satisfied based on available performance estimates. + +### Metric characteristics + +Returned metrics are: + +* **estimated**, not guaranteed +* intended for **planning decisions**, not runtime routing + +--- + +## Commercial viability + +The required capabilities are already supported or emerging in commercial and pre-commercial platforms: + +* Operator exposure platforms (e.g. NEF-based APIs) +* Edge computing platforms from hyperscalers and telecom vendors +* Aggregation platforms aligned with GSMA Open Gateway (e.g. Aduna-type models) + +These platforms already expose or internally use: + +* edge location information +* performance estimation +* application-aware routing or placement logic + +The proposal does not require new infrastructure, but rather a **standardized developer-facing abstraction** for planning. + +--- + +## YAML code available? + +NO + +--- + +## Validated in lab/productive environments? + +NO + +This proposal is currently at the **concept and API design stage**, motivated by identified gaps in the Edge Cloud API lifecycle. + +--- + +## Validated with real customers? + +NO + +The proposal is based on analysis of developer and platform requirements for edge deployment planning, particularly for latency-sensitive applications. + +--- + +## Validated with operators? + +NO + +The proposal has not yet been formally validated with operators and is submitted for Working Group discussion and evaluation. + +--- + +## Supporters in API Backlog Working Group + +To be added by the Working Group