Interface between Quantum and Classical Computers

Quantum Computing
Kubernetes
Cloud Computing
Observability
OpenTelemetry
MSc Thesis

MSc thesis and proof of concept, a Kubernetes-native control plane that treats quantum circuits and backends as managed, observable resources across the quantum–classical boundary.

Published

July 27, 2026

MSc in Quantum Computing & Quantum Technologies, Democritus University of Thrace.

Overview

Quantum computers are increasingly used not as standalone machines but as accelerators within hybrid workflows, in which a classical system prepares a circuit, submits it to a quantum processor, and interprets the probabilistic results it returns. This thesis treats the quantum–classical interface as a systems problem rather than a physics one. The orchestration, scheduling, and observability taken for granted in classical cloud-native operations are largely absent at the quantum execution boundary.

From a layered analysis of the interface and a survey of existing orchestration and observability work, the thesis derives a set of requirements and answers them with a proof of concept, the Quantum Circuit Controller (QCC).

Motivation

After a BSc in Applied Informatics and a dissertation on Cloud Computing (2012), I built my career in Systems and Site Reliability Engineering (SRE). In that discipline, orchestration, scheduling, and observability are foundational concerns, and their absence from quantum computing was conspicuous. The gap became apparent during the execution of Deutsch’s algorithm on IBM Quantum hardware. A two-qubit circuit incurred approximately three seconds of wall-clock time, whereas the gate operations themselves accounted for only microseconds.

The same behaviour recurred, more acutely, with Shor’s algorithm, where the selection of backend determined whether the computation yielded any usable signal. On one device the output was indistinguishable from noise. On another it survived. No mechanism existed to evaluate that selection a priori, so it reduced to manual inspection of calibration data and the submission of trial jobs. Addressing this gap, and treating quantum circuits and backends as first-class, observable resources, is the objective of this thesis.

Quantum Circuit Controller (QCC)

QCC is a control plane for the quantum–classical interface. It brings cloud-native orchestration and vendor-neutral, standards-based observability to quantum execution, so a backend becomes a managed, observable resource rather than an opaque remote endpoint.

Circuits and backends are modelled as Kubernetes custom resources, Circuit and QPU, under the qcc.io API group. A Go controller reconciles submitted circuits. A Python executor dispatches them through a provider adapter to either a simulator or IBM Quantum hardware. Execution emits OpenTelemetry metrics under a qcc.* namespace, exported through an OTel Collector to Prometheus and Grafana.

QCC system context: the qcc stack on Kubernetes, provider adapters to IBM Quantum and the Aer simulator, and the OpenTelemetry path to Prometheus and Grafana.

The design follows IBM’s Quantum-Centric Supercomputing (QCSC) reference architecture of Seelam et al. (2026), which identifies an orchestration and monitoring layer for hybrid workloads. QCC provides a cloud-native realisation of that layer on Kubernetes, occupying the System Orchestration layer and the management-and-monitoring cross-cut. It complements the Quantum Resource Management Interface (QRMI) of Bacher et al. (2025), the vendor-neutral middleware that treats a QPU as a schedulable resource. Where QRMI standardises provider-facing acquisition and execution beneath the scheduler, QCC models circuits and backends as first-class Kubernetes resources above it, and consumes such providers through adapters.

The design was exercised end-to-end on a representative quantum workload, across both a simulator and real IBM hardware. It satisfies the requirements set out in the thesis and makes the computational cost and behaviour of the interface visible. The thesis also proposes two standardisation pathways: a vendor-neutral execution interface and a set of quantum observability conventions.

Resources