The Function-as-a-Service (FaaS) paradigm offers a serverless approach that abstracts the management of underlying infrastructure, enabling developers to focus on application logic. However, leveraging infrastructure-aware features can further optimize serverless performance. We present a software prototype that enhances Apache OpenWhisk serverless platform with a novel architecture incorporating tAPP (topology-aware Allocation Priority Policies), a declarative language designed for specifying topology-aware scheduling policies. Through a case study involving distributed data access across multiple cloud regions, we show that tAPP can significantly reduce latency and minimizes performance variability compared to the standard OpenWhisk implementation.

Abstract is a stateful calculus in which clauses can be activated either through interactions with the external environment or by the evaluation of time expressions. Despite the apparent simplicity of its syntax and operational model, the combination of state evolution, time reasoning, and nondeterminism gives rise to significant analytical challenges. In particular, we show that determining whether a clause is never executed is undecidable. We formally prove that this undecidability result holds even for syntactically restricted fragments: namely, the time-ahead fragment, where all time expressions are strictly positive, the instantaneous fragment, where all time expressions evaluate to zero, and the determinate fragment, where the initial states of functions and events are disjoint. On the other hand, we identify a decidable subfragment: at the intersection of the instantaneous and determinate fragments reachability becomes decidable.

Cloud-Edge application deployment involves placing multiple software components on infrastructural topologies of heterogeneous nodes, ranging from Cloud servers to Internet-of-Things (IoT) edge devices. When multiple versions (or “ flavours ”) of a component are available, application managers must select a flavour for each deployed component, and assign these components to specific nodes, all while considering constraints such as dependencies, quality of service (QoS), budget, operational costs, and carbon emissions. In complex scenarios, finding the optimal deployment is often infeasible for human operators without automated tools to systematically explore the solution space. To address this challenge, we introduce FREEDA, a first constraint optimisation approach for deploying constrained and multi-flavoured applications on Cloud-Edge infrastructure topologies. We demonstrate the practical feasibility of FREEDA through experiments on a variety of realistic Cloud-Edge infrastructural topologies and component architectures. Furthermore, we benchmark FREEDA against Zephyrus, a comparable tool employing the same underlying solving technology. Empirical results show that FREEDA achieves strong scalability across a broad spectrum of realistic configurations and consistently outperforms Zephyrus.

We study a theory of asynchronous session types ensuring that well-typed processes terminate under a suitable fairness assumption. Fair termination entails starvation freedom and orphan message freedom namely that all messages, including those that are produced early taking advantage of asynchrony, are eventually consumed. The theory is based on a novel fair asynchronous subtyping relation for session types that is coarser than the existing ones. The type system is also the first of its kind that is firmly rooted in linear logic: fair asynchronous subtyping is incorporated as a natural generalization of the cut and axiom rules of linear logic and asynchronous communication is modeled through a suitable set of commuting conversions and of deep cut reductions in linear logic proofs.

International audience

Functions-as-a-Service (FaaS) is a Serverless Cloud paradigm where a platform manages the scheduling (e.g., resource allocation, runtime environments) of stateless functions. Recent work proposed using domain-specific languages to express per-function policies, e.g., policies that enforce the allocation on nodes that enjoy lower latencies to databases and services used by the function. Here, we focus on affinity-aware scenarios, i.e., where, for performance and functional requirements, the allocation of a function depends on the presence/absence of other functions on nodes. We present aAPP, an extension of a declarative, platform-agnostic language that captures affinity-aware scheduling at the FaaS level. We implement an aAPP-based prototype on Apache OpenWhisk. Besides proving that a FaaS platform can capture affinity awareness using aAPP and improve performance in affinity-aware scenarios, we use our prototype to show that aAPP imposes no noticeable overhead in scenarios without affinity constraints.

Serverless computing has extended its reach to encompass private edge cloud systems, aiming to enhance latency, security, and privacy while optimising resource usage. However, this extension comes with challenges such as running platforms and functions on disparate and resource-constrained devices. To respond to the challenges, we present FunLess, a Function-as-a-Service (FaaS) platform tailored for private edge cloud systems. Unlike conventional solutions relying on container technologies for function invocation, FunLess leverages WebAssembly (Wasm) as its runtime environment. This choice offers several advantages, including inherent security and isolation mechanisms crucial for data integrity and confidentiality, portability and consistent development and deployment, and a reduced memory footprint that allows functions to run on constrained edge devices.

Session types are widely used as abstractions of asynchronous message passing systems. Refinement for such abstractions is crucial as it allows improvements of a given component without compromising its compatibility with the rest of the system. In the context of session types, the most general notion of refinement is asynchronous session subtyping, which allows message emissions to be anticipated w.r.t. a bounded amount of message consumptions. In this paper we investigate the possibility to anticipate emissions w.r.t. an unbounded amount of consumptions: to this aim we propose to consider fair compliance over asynchronous session types and fair refinement as the relation that preserves it. This allows us to propose a novel variant of session subtyping that leverages the notion of controllability from service contract theory and that is a sound characterisation of fair refinement. In addition, we show that both fair refinement and our novel subtyping are undecidable. We also present a sound algorithm which deals with examples that feature potentially unbounded buffering. Finally, we present an implementation of our algorithm and an empirical evaluation of it on synthetic benchmarks.

State-of-the-art serverless platforms use hard-coded scheduling policies that are unaware of the possible topological constraints of functions. Considering these constraints when scheduling functions leads to sensible performance improvements, e.g., minimising loading times or data-access latencies. This issue becomes more pressing when considered in the emerging multi-cloud and edge-cloud-continuum systems, where only specific nodes can access specialised, local resources. To address this problem, we present a declarative language for defining serverless scheduling policies to express constraints on topologies of schedulers and execution nodes. We implement our approach as an extension of the OpenWhisk platform.

Abstract Session types are widely used as abstractions of asynchronous message passing systems. Refinement for such abstractions is crucial as it allows improvements of a given component without compromising its compatibility with the rest of the system. In the context of session types, the most general notion of refinement is the asynchronous session subtyping, which allows to anticipate message emissions but only under certain conditions. In particular, asynchronous session subtyping rules out candidates subtypes that occur naturally in communication protocols where, e.g., two parties simultaneously send each other a finite but unspecified amount of messages before removing them from their respective buffers. To address this shortcoming, we study fair compliance over asynchronous session types and fair refinement as the relation that preserves it. This allows us to propose a novel variant of session subtyping that leverages the notion of controllability from service contract theory and that is a sound characterisation of fair refinement. In addition, we show that both fair refinement and our novel subtyping are undecidable. We also present a sound algorithm, and its implementation, which deals with examples that feature potentially unbounded buffering.

Abstract We study the relationship between session types and behavioural contracts, representing Communicating Finite State Machines (CFSMs), under the assumption that processes communicate asynchronously. Session types represent a syntax-based approach for the description of communication protocols, while behavioural contracts, formally expressing CFSMs, follow an operational approach. We show the existence of a fully abstract interpretation of session types into a fragment of contracts that maps session subtyping into binary compliance-preserving CFSMs/behavioural contract refinement. In this way, on the one hand, we enrich the theory of session types with an operational characterization and, on the other hand, we use recent undecidability results for asynchronous session subtyping to obtain an original undecidability result for asynchronous CFSMs/behavioural contract refinement.

Microservices are highly modular and scalable Service Oriented Architectures. They underpin automated deployment practices like Continuous Deployment and Autoscaling. In this paper we formalize these practices and show that automated deployment — proven undecidable in the general case — is algorithmically treatable for microservices. Our key assumption is that the configuration life-cycle of a microservice is split into two phases: (i) creation, which entails establishing initial connections with already available microservices, and (ii) subsequent binding/unbinding with other microservices. To illustrate the applicability of our approach, we implement an automatic optimal deployment tool and compute deployment plans for a realistic microservice architecture, modeled in the Abstract Behavioral Specification (ABS) language.

We give two different notions of deadlock for systems based on active objects and futures. One is based on blocked objects and conforms with the classical definition of deadlock by Coffman, Jr. et al. The other one is an extended notion of deadlock based on blocked processes which is more general than the classical one. We introduce a technique to prove deadlock freedom of systems of active objects. To check deadlock freedom an abstract version of the program is translated into Petri nets. Extended deadlocks, and then also classical deadlock, can be detected via checking reachability of a distinct marking. Absence of deadlocks in the Petri net constitutes deadlock freedom of the concrete system.

The problem of deploying a complex software application has been formally investigated in previous work by means of the abstract component model named Aeolus. As the problem turned out to be undecidable, simplified versions of the model were investigated in which decidability was restored by introducing limitations on the ways components are described. In this paper, we take an opposite approach, and investigate the possibility to address a relaxed version of the deployment problem without limiting the expressiveness of the component model. We identify three problems to be solved in sequence: (i) the verification of the existence of a final configuration in which all the constraints imposed by the single components are satisfied, (ii) the generation of a concrete configuration satisfying such constraints, and (iii) the synthesis of a plan to reach such a configuration possibly going through intermediary configurations that violate the non-functional constraints.

Cloud computing offers the possibility to build sophisticated software systems on virtualized infrastructures at a fraction of the cost necessary just a few years ago. Nevertheless, the deployment of such complex systems is a serious issue due to the large number of involved software packages and services, and to their elaborated interdependencies. In this paper we address the challenge of automatizing this complex deployment process. We first formalize it as a planning problem and observe that standard planning tools can effectively solve it only on small and trivial instances. For this reason, we propose an ad hoc planning technique which we validate by means of a prototype implementation able to effectively solve this deployment problem also on instances of realistic size.

We explore the expressive power of languages that naturally model biochemical interactions relative to languages that only naturally model basic chemical reactions, identifying molecular association as the basic mechanism that distinguishes the former from the latter. We use a process algebra, the Biochemical Ground Form (BGF), that adds primitives for molecular association to CGF, which is a process algebra that has been proved to be equivalent to the traditional notations for describing basic chemical reactions. We first observe that, unlike CGF, BGF is Turing universal as it supports a finite precise encoding of Random Access Machines, which comprise a well-known Turing powerful formalism. Then we prove that the Turing universality of BGF derives from the interplay between the molecular primitives of association and dissociation. In fact, the elimination from BGF of the primitives already present in CGF does not reduce the computational strength of the process algebra, but if either association or dissociation is removed, BGF ceases to be Turing complete.

SOCK is a process calculus for the modeling of service oriented systems recently extended with primitives for dynamic fault and compensation handling. In this paper we investigate the relationships between the Sagas calculi for compensable flow composition and SOCK. First, we present an encoding of parallel Sagas (with interruption and centralized compensation) into SOCK. Then, we discuss a new semantics for parallel Sagas that we consider more adequate to the dynamic approach to fault and compensation handling.

Service Oriented Computing (SOC) allows for the composition of services which communicate using unidirectional one-way or bidirectional request-response communication patterns. Most service orchestration languages proposed so far provide also primitives for error handling based on fault, termination, and compensation handlers. Our work is motivated by the difficulties encountered in programming some error handling strategies using current error handling primitives. We propose as a solution an orchestration programming style in which handlers are dynamically installed. We assess our proposal by formalizing our approach as an extension of the process calculus SOCK and by proving that our formalization satisfies some expected high-level properties.

The investigation into the foundational aspects of linguistic mechanisms for programming long-running transactions (such as thescopeoperator of WS-BPEL) has recently renewed the interest in process algebraic operators that, due to the occurrence of a failure,interruptthe execution of one process, replacing it with another one called thefailure handler. We investigate the decidability of termination problems for two simple fragments of CCS (one with recursion and one with replication) extended with one of two such operators, theinterruptoperator of CSP and thetry-catchoperator for exception handling. More precisely, we consider the existential termination problem (existence of one terminated computation) and the universal termination problem (all computations terminate). We prove that, as far as the decidability of the considered problems is concerned, under replication there is no difference between interrupt and try-catch (universal termination is decidable while existential termination is not), while under recursion this is not the case (existential termination is undecidable while universal termination is decidable only for interrupt). As a consequence of our undecidability results, we show the existence of an expressiveness gap between a fragment of CCS and its extension with either the interrupt or the try-catch operator.

We investigate, in a process algebraic setting, a new notion of correctness for service compositions, which we callstrong service compliance: composed services are strong compliant if their composition is both deadlock and livelock free (this is the traditional notion of compliance), and whenever a message can be sent to invoke a service, it is guranteed to be ready to serve the invocation. We also define a new notion of refinement, calledstrong subcontract pre-order, suitable for strong compliance: given a composition of strong compliant services, we can replace any service with any other service in subcontract relation while preserving the overall strong compliance. Finally, we present a characterisation of the strong subcontract pre-order by resorting to the theory of a (should) testing pre-order.

In this paper we investigate the expressive power of three alternative approaches to the definition of infinite behaviours in process calculi, namely, recursive definitions, replication and iteration. We prove several results discriminating between the calculi obtained from a core CCS by adding the three mechanisms mentioned above. These results are derived by considering the decidability of four basic properties: termination (that is, all computations are finite); convergence (that is, the existence of a finite computation); barb (that is, the ability to perform an action on a given channel) and weak bisimulation. Our results, which are summarised in Table 1, show that the three calculi form a strict expressiveness hierarchy in that: all the properties mentioned are undecidable in CCS with recursion; only termination and barb are decidable in CCS with replication; all the properties are decidable in CCS with iteration. As a corollary, we also obtain a strict expressiveness hierarchy with respect to weak bisimulation, since there exist weak bisimulation preserving encodings of iteration in replication and of replication in recursion, whereas there are no weak bisimulation preserving encodings in the other directions.

The calculus of Mobile Ambients was proposed by Cardelli and Gordon as a foundational calculus for mobile computing. Since its introduction, the computational strength and the decidability of properties have been investigated for several fragments and variants of the standard calculus. We consider the problem of reachability and characterise a public (that is, restriction-free) fragment for which it is decidable. This fragment is obtained by removing the open capability and restricting the application of the replication operator to guarded processes only. This decidability result may appear surprising in combination with the fact that the same fragment was shown to be Turing complete by Maffeis and Phillips. Finally, we extend our decidability result in two ways: we first prove the decidability of a more general property called target reachability (according to which the target of interest for the reachability analysis consists of a possibly infinite set of processes) and then show that our decidability results also hold for a more general calculus, which includes the sophisticated communication mechanisms of Boxed Ambients, which is the most relevant variant of Mobile Ambients without the open capability.

We conclude this special issue ofMathematical Structures in Computer Sciencewith a list of Nadia Busi's scientific output (excluding the papers in the current volume).

In service oriented computing, choreography languages are used to specify multi-party service compositions. Two main approaches have been followed: the interaction-oriented approach of WS-CDL and the process-oriented approach of BPEL4Chor. We investigate the relationship between them.In particular, we consider several interpretations for interaction-oriented choreographies spanning from synchronous to asynchronous communication. Under each of these interpretations we characterize the class of interaction-oriented choreographies which have a process-oriented counterpart, and we formalize the notion of equivalence between the initial interaction-oriented choreography and the corresponding process-oriented one.

Peer-to-peer (P2P) has recently emerged as a promising model for supporting scalable networks composed of autonomous and spontaneously cooperating entities. The key concept in P2P is decentralization: the resources, the services, as well as the control are not in charge of specialized nodes in the network, but each node (called peer in this context) is directly involved in the management of all these aspects. Besides the advantages of decentralization (autonomy, adaptability, collaboration, and dinamicity just to mention few of them) one of the main drawbacks is the impossibility to predict the topology of the network, thus leaving at run-time any decision about the management of the interaction among the peers. For this reason, we consider useful to provide the developers of P2P applications with a high-level coordination language to be exploited to program the coordination among the peers. In this paper, we present [Formula: see text], a new data-driven coordination model suitable for P2P networks, and we describe [Formula: see text], an implementation of the [Formula: see text] coordination model based on the JXTA peer-to-peer technology.