Jacopo Mauro

Automatically deriving executable and efficient constraint programming models from natural language descriptions has long been a central goal in optimisation and constraint programming. Large Language Models (LLMs) appear to bring this vision closer, showing promising results in generating models for classical benchmark problems. However, part of this apparent success may stem from data contamination rather than genuine reasoning capabilities, as many standard CP problems are likely represented in the training data of modern LLMs. In this initial investigation, we study the robustness of LLM-based constraint modelling by rephrasing and perturbing a set of 10 constraint problems while preserving their underlying structure. The modifications alter contextual elements and introduce misleading or distracting information to explore whether models rely on semantic understanding or superficial textual cues. We compare the models generated by three representative LLMs across the original and modified problem descriptions. Our qualitative analysis shows that, although LLMs frequently produce syntactically valid and semantically plausible models, their performance degrades significantly under contextual and linguistic variation, suggesting strong sensitivity to wording and contextual cues.

APP is a declarative language for the definition of custom function scheduling on the worker nodes available in serverless Function as a Service (FaaS) platforms. Current APP platforms assume a central control point that users can access to issue the execution of functions. We propose dAPP, a decentralised version of APP where multiple control points tackle both scaling and resilience issues of existing implementations. We formalise the semantics of dAPP and use it to show that despite APP and dAPP are not equivalent according to a trace-based semantics, they produce the same distributions of functions over the available workers. We also present an implementation of dAPP using the FunLess FaaS platform – tailored for private edge-cloud and multi-cloud environments – and show initial empirical evidence of performance improvements in setups where both the platform and function invocations are spread across multiple locations.

One of the long-standing goals in optimisation and constraint programming is to describe a problem in natural language and automatically obtain an executable, efficient model. Large language models appear to bring this vision closer, showing impressive results in automatically generating models for classical benchmarks. However, much of this apparent success may derive from data contamination rather than genuine reasoning: many standard CP problems are likely included in the training data of these models. To examine this hypothesis, we systematically rephrased and perturbed a set of well-known CSPLib problems to preserve their structure while modifying their context and introducing misleading elements. We then compared the models produced by three representative LLMs across original and modified descriptions. Our qualitative analysis shows that while LLMs can produce syntactically valid and semantically plausible models, their performance drops sharply under contextual and linguistic variation, revealing shallow understanding and sensitivity to wording.

Distributed applications employ Kubernetes for scalable, fault-tolerant deployments over computer clusters, where application components run in groups of containers called pods. The scheduler, at the heart of Kubernetes' architecture, determines the placement of pods given their priority and resource requirements on cluster nodes. To quickly allocate pods, the scheduler uses lightweight heuristics that can lead to suboptimal placements and resource fragmentation, preventing allocations of otherwise deployable pods on the available nodes. We propose the usage of constraint programming to find the optimal allocation of pods satisfying all their priorities and resource requests. Implementation-wise, our solution comes as a plug-in to the default scheduler that operates as a fallback mechanism when some pods cannot be allocated. Using the OR-Tools constraint solver, our experiments on small-to-mid-sized clusters indicate that, within a 1-second scheduling window, our approach places more higher-priority pods than the default scheduler (possibly demonstrating allocation optimality) in over 44\% of realisable allocation scenarios where the default scheduler fails, while certifying that the default scheduler's placement is already optimal in over 19\% of scenarios. With a 10-second window, our approach improves placements in over 73\% and still certifies that the default scheduler's placement is already optimal in over 19\% of scenarios.

FunLess is a Function-as-a-Service (FaaS) platform tailored for private edge cloud systems. FunLess leverages WebAssembly as its runtime environment for performance, function isolation, and support for heterogeneous devices, crucial for extending the coverage of serverless computing to private edge cloud systems. We benchmark FunLess against three production-ready, widely adopted open-source FaaS platforms—OpenFaaS, Fission, and Knative—under different deployment scenarios, characterised by the presence/absence of constrained-resource devices (Raspberry Pi 3B+) and the (in)accessibility of container orchestration technologies—Kubernetes. Our results confirm that FunLess is a suitable solution for FaaS private edge cloud systems since it achieves performance comparable to the considered FaaS alternatives while it is the only fully-deployable alternative on constrained-resource devices.