Synthetic data has emerged as a promising avenue for privacy-preserving data sharing. However, constructing synthetic data generators necessitates access to the real dataset, posing challenges, particularly when data features are disparately distributed across different organizations. Vertical Federated Learning (VFL) is a collaborative approach to training machine learning models among distinct tabular data holders, such as financial institutions, who possess disjoint features for the same group of customers. In this paper, we introduce the GTV framework for Generating Tabular Data via Vertical Federated Learning and demonstrate that VFL can be successfully used to implement GANs for distributed tabular data in a privacy-preserving manner, with performance close to centralized GANs which assume shared data. We make design choices with respect to the distribution of GAN generator and discriminator models, and we introduce a training-with-shuffling technique so that no party can reconstruct training data from the GAN conditional vector. The paper presents (1) an implementation of GTV, (2) a detailed quality evaluation of GTV-generated synthetic data, (3) an examination of the GTV framework for different data distributions and number of clients, and (4) an analysis of GTV’s robustness against Membership Inference Attacks with different settings of Differential Privacy, for a range of datasets with diverse distribution characteristics. Our results demonstrate that GTV can consistently generate high-fidelity synthetic tabular data of comparable quality to that generated by a centralized GAN algorithm. The difference in machine learning utility can be as low as 2.7%, even under extremely imbalanced data distributions across clients. Code is available at: https://github.com/zhao-zilong/gtv

Collaborative blockchain storage is one of the solutions to the blockchain storage scalability challenge, allowing individual nodes to store partial blockchain data. Collaborative blockchains give up redundancy for scalability. However, redundancy is a determinant of blockchain’s decentralization and data availability. In this article, we study the possibility of balancing the scalability–redundancy tradeoff in collaborative blockchains. We first present a model for the evaluation of a data distribution scheme for its efficiency in terms of redundancy. Then, we present a fully decentralized self-regulated data distribution scheme assuring storage scalability and redundancy for collaborative blockchains based on fully redundant global state. The proposed solution does not incur additional communication. We examine the efficiency of the proposed scheme based on experiments. Our results demonstrate that the proposed system assures a level of redundancy for a given set of nodes and data volume, optimizes storage space utilization, and provides uniform distribution of data.

In this research, we investigate the influence of utilizing artificial intelligence (AI)-generated photographs on malicious bots that engage in disinformation, fraud, reputation manipulation, and other types of malicious activity on social networks. Our research aims to compare the performance metrics of social bots that employ AI photos with those that use other types of photographs. To accomplish this, we analyzed a dataset with 13 748 measurements of 11 423 bots from the VK social network and identified 73 cases where bots employed generative adversarial network (GAN)-photos and 84 cases where bots employed diffusion or transformers photos. We conducted a qualitative comparison of these bots using metrics such as price, survival rate, quality, speed, and human trust. Our study findings indicate that bots that use AI-photos exhibit less danger and lower levels of sophistication compared to other types: AI-enhanced bots are less expensive, less popular on exchange platforms, of inferior quality, less likely to be operated by humans, and, as a consequence, faster and more susceptible to being blocked by social networks. We also did not observe any significant difference between GAN-based and diffusion/transformers-based bots, indicating that diffusion/transformers models did not contribute to increased bot sophistication compared to GAN models. Our contributions include a proposed methodology for evaluating the impact of photos on bot sophistication, along with a publicly available dataset for other researchers to study and analyze bots. Our research findings suggest a contradiction to theoretical expectations: in practice, bots using AI-generated photos pose less danger.

The rise in adoption of cryptoassets has brought many new and inexperienced investors in the cryptocurrency space. These investors can be disproportionally influenced by information they receive online, and particularly from social media. This paper presents a dataset of crypto-related bounty events and the users that participate in them. These events coordinate social media campaigns to create artificial "hype" around a crypto project in order to influence the price of its token. The dataset consists of information about 15.8K cross-media bounty events, 185K participants, 10M forum comments and 82M social media URLs collected from the Bounties(Altcoins) subforum of the BitcoinTalk online forum from May 2014 to December 2022. We describe the data collection and the data processing methods employed and we present a basic characterization of the dataset. Furthermore, we discuss potential research opportunities afforded by the dataset across many disciplines and we highlight potential novel insights into how the cryptocurrency industry operates and how it interacts with its audience.

Federated Learning (FL) has emerged as a potentially powerful privacy-preserving machine learning method-ology, since it avoids exchanging data between participants, but instead exchanges model parameters. FL has traditionally been applied to image, voice and similar data, but recently it has started to draw attention from domains including financial services where the data is predominantly tabular. However, the work on tabular data has not yet considered potential attacks, in particular attacks using Generative Adversarial Networks (GANs), which have been successfully applied to FL for non-tabular data. This paper is the first to explore leakage of private data in Federated Learning systems that process tabular data. We design a Generative Adversarial Networks (GANs)-based attack model which can be deployed on a malicious client to reconstruct data and its properties from other participants. As a side-effect of considering tabular data, we are able to statistically assess the efficacy of the attack (without relying on human observation such as done for FL for images). We implement our attack model in a recently developed generic FL software framework for tabular data processing. The experimental results demonstrate the effectiveness of the proposed attack model, thus suggesting that further research is required to counter GAN-based privacy attacks.

In proof-of-work based blockchains such as Ethereum, verification of blocks is an integral part of establishing consensus across nodes. However, in Ethereum, miners do not receive a reward for verifying. This implies that miners face the Verifier's Dilemma: use resources for verification, or use them for the more lucrative mining of new blocks? We provide an extensive analysis of the Verifier's Dilemma, using a data-driven model-based approach that combines closed-form expressions, machine learning techniques and discrete-event simulation. We collect data from over 300,000 smart contracts and experimentally obtain their CPU execution times. Gaussian Mixture Models and Random Forest Regression transform the data into distributions and inputs suitable for the simulator. We show that, indeed, it is often economically rational not to verify, in particular for miners with less hashing power. We consider two approaches to mitigate the implications of the Verifier's Dilemma, namely parallelization and active insertion of invalid blocks, both will be shown to be effective.

Attack graphs are commonly used to analyse the security of medium-sized to large networks. Based on a scan of the network and likelihood information of vulnerabilities, attack graphs can be transformed into Bayesian Attack Graphs (BAGs). These BAGs are used to evaluate how security controls affect a network and how changes in topology affect security. A challenge with these automatically generated BAGs is that cycles arise naturally, which make it impossible to use Bayesian network theory to calculate state probabilities. In this paper we provide a systematic approach to analyse and perform computations over cyclic Bayesian attack graphs. We present an interpretation of Bayesian attack graphs based on combinational logic circuits, which facilitates an intuitively attractive systematic treatment of cycles. We prove properties of the associated logic circuit and present an algorithm that computes state probabilities without altering the attack graphs (e.g., remove an arc to remove a cycle). Moreover, our algorithm deals seamlessly with any cycle without the need to identify their type. A set of experiments demonstrates the scalability of the algorithm on computer networks with hundreds of machines, each with multiple vulnerabilities.

Summary The two‐phase commit (2PC) protocol has long been known to have a provably inevitable vulnerability to blocking or non‐progress amidst server crashes, even when the distributed database system guarantees the most demanding timing‐related or “synchrony” requirements. Our aim here is to eliminate this vulnerability by using a blockchain for coordinating 2PC execution. We present the impossibilities, the possibilities, the cost, and the trade‐offs in this blockchain‐based approach to blocking‐free management of distributed transactions. We prove that a non‐blocking and blockchain‐coordinated 2PC protocol can exist only if both the blockchain and distributed database systems meet synchrony requirements; otherwise, although blocking remains eliminated, transactions can unnecessarily abort. We present a blockchain‐coordinated 2PC protocol and provide rigorous arguments for its correctness under the synchrony requirements. We then implement this protocol on the Ethereum Testnet and demonstrate, through our experiments, that the monetary cost of executing smart contracts is quite small, that the protocol performance slows down when using a public blockchain like Ethereum, and that even major violations of synchrony requirements lead only to relatively small increases in unnecessary aborts. We thus identify a trade‐off between improving protocol performance and admitting a risk that transactions could occasionally abort unnecessarily.

The Internet of Things (IoT) is becoming a backbone of sensing infrastructure to several mission-critical applications such as smart health, disaster management, and smart cities. Due to resource-constrained sensing devices, IoT infrastructures use Edge datacenters (EDCs) for real-time data processing. EDCs can be either static or mobile in nature, and this article considers both of these scenarios. Generally, EDCs communicate with IoT devices in emergency scenarios to evaluate data in real-time. Protecting data communications from malicious activity becomes a key factor, as all the communication flows through insecure channels. In such infrastructures, it is a challenging task for EDCs to ensure the trustworthiness of the data for emergency evaluations. The current communication security pattern of “communication before authentication” leaves a “black hole” for intruders to become part of communication processes without authentication. To overcome this issue and to develop security infrastructures for IoT and distributed Edge datacenters, this article proposes a user-centric security solution. The proposed security solution shifts from a network-centric approach to a user-centric security approach by authenticating users and devices before communication is established. A trusted controller is initialized to authenticate and establishes the secure channel between the devices before they start communication between themselves. The centralized controller draws a perimeter for secure communications within the boundary. Theoretical analysis and experimental evaluation of the proposed security model show that it not only secures the communication infrastructure but also improves the overall network performance.

Ethereum is a public (permissionless) blockchain with a Turing complete execution machine for smart contracts. Miners that execute a smart contract receive a fee determined by the gas associated with the operation codes (opcodes) in the smart contract. It is important that the gas award is proportional to the computation resources required, to assure that incentives are aligned and denial of service attacks are avoided. Currently, the amount of gas awarded is set statically for each opcode, but it is unknown if these values are correct for various computer architectures. Therefore, we propose in this paper a benchmark approach to assess the computational resources required per opcode. We apply the benchmark approach to PC and MAC as a first illustration of the approach.

Blockchain is a highly popular paradigm for non-centralized applications, especially in finance and trade. Performance is a major challenge for blockchains, since consensus approaches are known not to scale. In this presentation we address blockchain performance, from the perspective of model-based prediction as well as benchmark-based assessment. We present research results about smart contracts in the Ethereum blockchain and discuss the requirements for generic benchmarks for blockchain performance. Benchmarking is a common approach to compare industry-class systems. As blockchain technologies mature, the role of reliable benchmarks will become increasingly important. However, definitions of benchmarks for blockchains are still in their infancy. We argue that there is a clear need for benchmarks, and that benchmarks should be based on the sound scientific principles of metrology [1]. A variety of important performance issues should be addressed, including the performance of the proof (be it work, stake, or other), transaction processing and block creation. Moreover, in all these situations, establishing energy consumption benchmarks is critical in determining if incentives are in place for miners to operate the blockchain system. A particularly interesting element in some blockchains is the mechanism of smart contracts. For instance, in Ethereum, the fees associated with executing contracts depend on the benchmarked performance of the operation code. In [2] it was demonstrated that uncertainty with respect to the correctness of the anticipated execution time impacts the decisions miners will take. We will discuss improved benchmarking approaches for operational code.

Cloud computing has become an irreversible trend. Together comes the pressing need for verifiability, to assure the client the correctness of computation outsourced to the cloud. Existing verifiable computation techniques all have a high overhead, thus if being deployed in the clouds, would render cloud computing more expensive than the on-premises counterpart. To achieve verifiability at a reasonable cost, we leverage game theory and propose a smart contract based solution. In a nutshell, a client lets two clouds compute the same task, and uses smart contracts to stimulate tension, betrayal and distrust between the clouds, so that rational clouds will not collude and cheat. In the absence of collusion, verification of correctness can be done easily by crosschecking the results from the two clouds. We provide a formal analysis of the games induced by the contracts, and prove that the contracts will be effective under certain reasonable assumptions. By resorting to game theory and smart contracts, we are able to avoid heavy cryptographic protocols. The client only needs to pay two clouds to compute in the clear, and a small transaction fee to use the smart contracts. We also conducted a feasibility study that involves implementing the contracts in Solidity and running them on the official Ethereum network.

In this practical experience report we discuss a court case in which one of the authors was expert witness. This UK civil case considered possible fraud in an online product promotion competition, with participants being denied prizes because they were considered to have cheated. The discussion in this paper aims to provide a practice-led perspective on the link between technology and legal issues in the design of online games and web applications. The paper presents the court's questions and the witness responses, and also provides a synopsis of analysis of data in the web server log file presented to court. Based on the insights gained, we present guidelines for the design of online competitions and for client-server web applications implementing it. As we will see, the case turned out to be about design of socio-technical systems, not about advanced technologies. It illustrates the need to identify practically relevant threat models and pragmatic security solutions that balance business, legal and usability concerns.

Continuous authentication is receiving increased attention from providers of on-line services, particularly due to the ability of mobile apps to collect user-specific sensor data. However, the approaches proposed so far are either not accurate enough to provide a high-quality user experience or restricted by engineering challenges to capture data continuously. In this paper, we propose an approach based on a deep learning autoencoder, which achieves an equal error rate as low as 2:2% in tested real-world scenarios. The suggested system only relies on accelerometer data and does not require a high number of features, therefore reducing the computational burden. We discuss the balance between the number of dimensional features and the re-authentication time, which decreases as the number of dimensions increases. We also discuss parameter selection for real-world scenarios e.g. depth of the architecture, time elapsed before re-building the model and length of the training dataset and possible approaches to find the optimal trade-off between accuracy and usability required for each particular context.

In this paper we present an attack, which allows fraudulent transactions to be collected from EMV contactless credit and debit cards without the knowledge of the cardholder. The attack exploits a previously unreported vulnerability in EMV protocol, which allows EMV contactless cards to approve unlimited value transactions without the cardholder's PIN when the transaction is carried out in a foreign currency. For example, we have found that Visa credit cards will approve foreign currency transactions for any amount up to ∈999,999.99 without the cardholder's PIN, this side-steps the £20 contactless transaction limit in the UK. This paper outlines our analysis methodology that identified the flaw in the EMV protocol, and presents a scenario in which fraudulent transaction details are transmitted over the Internet to a "rogue merchant" who then uses the transaction data to take money from the victim's account. In reality, the criminals would choose a value between ∈100 and ∈200, which is low enough to be within the victim's balance and not to raise suspicion, but high enough to make each attack worthwhile. The attack is novel in that it could be operated on a large scale with multiple attackers collecting fraudulent transactions for a central rogue merchant which can be located anywhere in the world where EMV payments are accepted.

While security algorithms are utilized to protect system resources from misuse, using a single algorithm such as CAPTCHAs and Spam-Filters as a defence mechanism can work to protect a system against current attacks. However, as attackers learn from their attempts, this algorithm will eventually become useless and the system is no longer protected. We propose to look at a set of algorithms as a combined defence mechanism to maximize the time taken by attackers to break a system. When studying sets of algorithms, diverse issues arise in terms of how to construct them and in which order or in which combination to release them. In this paper, we propose a model based on Stochastic Petri Nets, which describe the interaction between an attacker, the set of algorithms used by a system, and the knowledge gained by the attacker with each attack. In particular, we investigate the interleaving of dependent algorithms, which have overlapping rules, with independent algorithms, which have a disjoint set of rules. Based on the proposed model, we have analyzed and evaluated how the order can impact the time taken by an attacker to break a set of algorithms. Given the mean time to security failure (MTTSF) for a system to reach a failure state, we identify an improved approach to the release order of a set of algorithms in terms of maximizing the time taken by the attacker to break them. Further, we show a prediction of the attacker's knowledge acquisition progress during the attack process.

An intrusion and attack detection system usually focuses on classifying a record as either normal or abnormal. In some cases such as insider attacks, attackers rely on feedback from the attacked system, which enables them to gradually manipulate their attempts in order to avoid detection. This paper proposes the notion of accumulative manipulation that can be observed through a number of attempts accomplished by the attacker, which forms the basis of the Attacker Learning Curve (ALC). Based on a controlled experiment, we first show that the ALC for three different attack strategies are consistent between two different groups of subjects. We then define a strategy detection mechanism, which is experimentally shown to be accurate more than 70% of the time.

A critical challenge in cloud computing is assuring confidentiality and integrity for the execution of arbitrary software in a consumer's virtual machine. The problem arises from having multiple virtual machines sharing hardware resources in the same physical host. A security critical resource is random access memory, which in the current version of the Xen hyper visor is vulnerable to attacks. Like previous work demonstrated, this vulnerability originates from Xen adopting avery permissive memory access model for its management virtual machine (Dom0). The model assumes it is safe to grant Dom0full access to the memory space allocated to consumer's virtual machines. In this paper, we first present a sophisticated attack which makes it possible to compromise security-sensitive information resident in the memory area of a particular process executing in a virtual machine. The attack demonstration consists in subverting the new inter-virtual machine communication mechanism, libvchan, which is under development for the Xen hyper visor. This attack allows us to propose and implement a proof of concept for a lightweight mandatory memory access control mechanism for Xen, which achieves a better overall memory access model forDom0. We then propose an architecture which takes advantage of our memory protection mechanism and previous work to achievedefense in depth in cloud computing.

Probabilistic and stochastic models are routinely used in performance, dependability and security evaluation, and determining appropriate values for model parameters is a long-standing problem in the practical use of such models. With the increasing emphasis on human aspects and business considerations, data collection to estimate parameter values often gets prohibitively expensive, since it may involve questionnaires, costly audits or additional monitoring and processing. In this paper we articulate a set of optimization problems related to data collection, and provide efficient algorithms to determine the optimal data collection strategy for a model. The main idea is to model the uncertainty of data sources and determine its influence on output accuracy by solving the model. This approach is particularly natural for data sources that rely on sampling, such as questionnaires or monitoring, since uncertainty can be expressed using the central limit theorem. We pay special attention to the efficiency of our optimization algorithm, using ideas inspired by importance sampling to derive optimal strategies for a range of parameter values from a single set of experiments.

Recent advances in the research of usable security have produced many new security mechanisms that improve usability. However, these mechanisms have not been widely adopted in practice. In most organisations, IT security managers decide on security policies and mechanisms, seemingly without considering usability. IT security managers consider risk reduction and the business impact of information security controls, but not the impact that controls have on users. Rather than trying to remind security managers of usability, we present a new paradigm -- a stealth approach which incorporates the impact of security controls on users' productivity and willingness to comply into business impact and risk reduction. During two 2-hour sessions, 3 IT security managers discussed with us mock-up tool prototypes that embody these principles, alongside a range of potential usage scenarios (e.g. cloud-based password-cracking attacks and "hot-desking" initiatives). Our tool design process elicits findings to help develop mechanisms to visualise these tradeoffs.

Cloud based infrastructures are rapidly becoming a destination of choice to host a variety of applications ranging from high availability enterprise services and online TV stations, to batch oriented scientific computations. With investments of billions of dollars, the fortunes of dozens of companies, and major research initiatives staked on its success, it is clear that cloud computing is here to stay. However, it is not yet clear whether cloud services can be a dependable alternative to dedicated infrastructure. On one hand, availability and privacy are serious challenges for applications hosted on cloud infrastructure. On the other hand, a cloud provider's economies of scale allow levels of investment in redundancy and dependability that are difficult to match for smaller operators. Furthermore, the ability to monitor large numbers of applications can enable ldquowisdom of crowdsrdquo approaches to provide enhanced security much in the same way that network providers have been able to do with worms and DDoS attacks. The panel will discuss new dependability related challenges and opportunities that arise in the context of cloud computing, some examples of which are as follows.

Security managers often regard human behaviour as a security liability, but they should accommodate it within their organisation's information security management procedures. To further the comprehension of human-behavioural factors we develop an information security ontology. This ontology is intended for organisations that aim to maintain compliance with external standards (in this case ISO27002) while considering the security behaviours of individuals within the organisation.

Layer 2 traceback is an important component of end-to-end packet traceback. Whilst IP traceback identifies the origin network, L2 traceback extends the process to provide a more fine-grained result. Other known proposals have exposed the difficulties of L2 traceback in switched Ethernet. We build on our earlier work and improve in a number of dimensions. Memory requirements are decreased by maintaining 'connection records' rather than logging all frames. Our switchport resolution algorithm provides error detection by correlating MAC address table values from two adjacent switches. Our solution also takes stock of potential transformations to packet data as this leaves the local network. We have implemented the core algorithm and used data from available WAN traces to demonstrate the potential memory efficiency of our approach.

This workshop summary gives a brief overview of the workshop on ldquoResilience Assessment and Dependability Benchmarkingrdquo held in conjunction with the 38th IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2008). The workshop aims at the presentation and exchange of ideas from the world wide research community and fostering discussions in order to give answers to the need for improving trustworthiness and understand the current risks inherent to computer systems and infrastructures. In particular the workshop aims at addressing key research challenges related to effective and accurate methods for measuring, assessing and benchmarking dependability and resilience.

This paper describes work to ease the resource allocation problem in the domain of game server hosting. A solution was sought that required no alteration to game server code and would not inhibit a player's gaming experience. Although an academic work, the problem is tackled in a commercial setting.

Self-adaptive security offers great potential in providing timely and fine grained security control. In this paper we experimentally investigate the effects of a security adaptation in various client-server scenarios. We do this using SSL/TLS and show how client load patterns determine the overall performance effect of a security adaptation. We also demonstrate the viability of such a system using adaptive SSL. Experiments and analysis show that client arrival rate is not sufficient for making an adaptation decision. The average requested file size and the client session duration play key roles in the performance effect of such an adaptation. Adaptation with file sizes larger than 8192 bytes and small session durations result in greater performance impact when the server is under heavy load.

Restarts or retries are a common phenomenon in computing systems, for instance, in preventive maintenance, software rejuvenation, or when a failure is suspected. Typically, one sets a time-out to trigger the restart. We analyze and optimize time-out strategies for scenarios in which the expected required remaining time of a task is not always decreasing with the time invested in it. Examples of such tasks include the download of Web pages, randomized algorithms, distributed queries, and jobs subject to network or other failures. Assuming the independence of the completion time of successive tries, we derive computationally attractive expressions for the moments of the completion time, as well as for the probability that a task is able to meet a deadline. These expressions facilitate efficient algorithms to compute optimal restart strategies and are promising candidates for pragmatic online optimization of restart timers

Increasingly frequently, Grid software is being proposed as a software platform for operational management of enterprise computing systems, including desktops, data centres and business applications. We argue in this paper that the challenge for future management software is ‘enabling automated and business-driven management, for systems with increasing scale, heterogeneity and federation challenges’. Among the currently available middlewares, the Grid platform is undoubtedly best positioned to respond to this challenge, and we review several recent technologies that utilize Grid software for management purposes. However, Grid software also has its limitations. Its software architecture provides no particular support to building self-managing distributed systems and is therefore most viable for traditional hierarchical management systems, providing rudimentary self-management capabilities through control loops.

Grid computing has relied on "best effort" as the guiding principal of operation. However, commercial grids need to provide much stricter guarantees. These guarantees have to be specified in terms of service level agreements and have to be monitored and assured. We propose an architecture for specifying and monitoring service level agreements to achieve the above. The architecture relies on a network of communicating proxies each maintaining SLAB committed within the administrative domain of the proxy. SLAB are either negotiated between or specified to management proxies, and they are responsible for automated monitoring of data and for triggering evaluations of the registered SLAs. An unambiguous and,flexible language for formalizing SLAB is presented to achieve the above. The management proxy allows reasoning about the overall status of SLAB related to an application context across multiple administrative domains by contacting and querying involved management proxies, obtaining measurement information from multiple proxies, if needed and performing a consolidated SLA evaluation. The process of measurement collection, and SLA evaluation is automated and based on web services technology. The scenario considers HP Utility Data Center as a typical Commercial Grid deployment environment.

A component-based software development approach has become a trend in integrating modern software systems. To ensure the overall reliability of an integrated software system, its software components have to meet certain reliability requirements, subject to some testing schedule and resource constraints. Efficiency improvement of the system-testing can be formulated as a combinatorial optimization problem with known cost, reliability, effort and other attributes of the system components. This paper considers "software component testing resource allocation" for a system with single or multiple applications, each with a pre-specified reliability requirement. The relation between failure rates of components and "cost to decrease this rate" is modeled by various types of reliability-growth curves. Closed-form solutions to the problem for systems with one single application are developed, and then "how to solve the multiple application problem using nonlinear programming techniques" are described. Also examined are the interactions between the system components, and inter-component failure dependencies are included in the modeling formula. In addition to regular systems, the technique is extended to address fault-tolerant systems. A procedure for a systematic approach to the testing resource allocation problem is developed, and its application in a case study of a telecommunications software system is described. This procedure is automated in a reliability allocation tool for an easy specification of the problem and an automatic application of the technique. This methodology gives the basic approach to optimization of testing schedules, subject to reliability constraints. This adds "interesting new optimization opportunities in the software testing phase" to the existing optimization literature that is concerned with structural optimization of the software architecture. Merging these two approaches improves the reliability planning accuracy in component-based software development.

We present the quality-of-service query service (QQS) a CORBA service which provides quality-of-service-related "decisions" for the configuration and parameterization of CORBA applications. Applications and services query QQS for those QoS-related decisions, and input data into QQS, so that evaluators have the necessary information to execute their decision-making algorithms. QQS consists of a database, a registration, naming and dramatic-loading facility for evaluators, and a registration and dynamic-loading facility for data collectors. The data model underlying the database is a QoS-oriented data model specially designed for QQS, such that a maximum amount of relevant queries and evaluators fit into QQS. This paper describes the motivation, design, implementation and use of the QoS query service.

The phenomenon of software aging refers to the accumulation of errors during the execution of the software which eventually results in it's crash/hang failure. A gradual performance degradation may also accompany software aging. Pro-active fault management techniques such as "software rejuvenation" (Y. Huang et al., 1995) may be used to counteract aging if it exists. We propose a methodology for detection and estimation of aging in the UNIX operating system. First, we present the design and implementation of an SNMP based, distributed monitoring tool used to collect operating system resource usage and system activity data at regular intervals, from networked UNIX workstations. Statistical trend detection techniques are applied to this data to detect/validate the existence of aging. For quantifying the effect of aging in operating system resources, we propose a metric: "estimated time to exhaustion", which is calculated using well known slope estimation techniques. Although the distributed data collection tool is specific to UNIX, the statistical techniques can be used for detection and estimation of aging in other software as well.

To ensure an overall reliability of an integrated software system, software components of the system have to meet certain reliability requirements, subject to some testing schedule and resource constraints. The system testing activity can be formulated as a combinatorial optimization problem with known cost, reliability, effort and other attributes of the system components. In this paper, we consider the software component reliability allocation problem for a system with multiple applications. The failure rate of components used to build the applications are related to the testing cost through various types of reliability growth curves. We achieve closed-form solutions to problems where there is one single application in the system. Analytical solutions are not readily available when there are multiple applications; however, numerical solutions can be obtained using a nonlinear programming tool. To ease the specification of the optimization problem, we develop a GUI front-end to existing mathematical software. We present a systematic outline of the problem formulation and solution, and apply this to an example of a telecommunication software system.

Stochastic Petri nets (SPNs) and extensions are a popular method for evaluating a wide variety of systems. In most cases, their numerical solution requires generating a state-level stochastic process, which captures the behavior of the SPN with respect to a set of specified performance measures. These measures are commonly defined at the net level by means of a reward variable. In this paper, we discuss issues regarding the generation of state-level reward models for systems specified as stochastic activity networks (SANs) with "step-based reward structures". Step-based reward structures are a generalization of previously proposed reward structures for SPNs and can represent all reward variables that can be defined on the marking behavior of a net. While discussing issues related to the generation of the underlying state-level reward model, we provide an algorithm to determine whether a given SAN is "well-specified" A SAN is well-specified if choices about which instantaneous activity completes among multiple simultaneously-enabled instantaneous activities do not matter, with respect to the probability of reaching next possible stable markings and the distribution of reward obtained upon completion of a timed activity. The fact that a SAN is well specified is both a necessary and sufficient condition for its behavior to be completely probabilistically specified, and hence is an important property to determine.

The ability of error recovery mechanisms to make the Fiber Distributed Data Interface (FDDI) satisfy real-time performance constraints in the presence of errors is analyzed. A complicating factor in these analyses is the rarity of the error occurrences, which makes direct simulation unattractive. Therefore, a fast simulation technique, called injection simulation, which makes it possible to analyze the performance of FDDI, including its fault tolerance behavior, was developed. The implementation of injection simulation for polling models of FDDI is discussed, along with simulation results.>