Jun Ho Huh

Hybrid spam filtering for mobile communication

Spam messages are an increasing threat to mobile communication. Several mitigation techniques have been proposed, including white and black listing, challenge-response and content-based filtering. However, none are perfect and it makes sense to use a combination rather than just one. We propose an anti-spam framework based on the hybrid of content-based filtering and challenge-response. A message, that has been classified as uncertain through content-based filtering, is checked further by sending a challenge to the message sender. An automated spam generator is unlikely to send back a correct response, in which case, the message is classified as spam. Our simulation results show the trade-off between the accuracy of anti-spam classifiers and the incurring traffic overhead, and demonstrate that our hybrid framework is capable of achieving high accuracy regardless of the content-based filtering algorithm being used.

Trusted Logging for Grid Computing

The rise of many kinds of grid systems, and associated security threats, makes very necessary the provision of trustworthy services for audit and logging. However, existing solutions tend to put little emphasis on the security of logging. We present a number of use cases where the logs have security properties in their own rights, and so the logs themselves are highly privileged: hence, these need to be integrity and confidentiality protected while being generated, accessed, reconciled and analysed with distributed services spanning across multiple administrative domains. We derive a common set of secure logging requirements to address the security gaps which exist between these use cases and existing solutions.From the requirements, we propose a novel logging architecture for the grid based on virtual machine (VM) isolation and trusted computing capabilities: a small number of privileged driver VMs trigger all trusted logging requests and forward them to the secure logging service running within the log security manager. The logging service verifies the integrity of the log data and the security configurations of these driver VMs (log generators) before storing the logs.

Trustworthy Log Reconciliation for Distributed Virtual Organisations

Secure management of logs in an organisational grid environment is often considered a task of low priority. However, it must be rapidly upgraded when the logs have security properties in their own right. We present several use cases where log integrity and confidentiality are essential, and propose a log reconciliation architecture in which both are ensured. We use a combination of trusted computing and virtualization to enable blind log analysis , allowing users to see the results of legitimate queries, while still withholding access to privileged raw data.

Managing application whitelists in trusted distributed systems

Many distributed batch systems, such as computational grids, require a level of integrity protection to guarantee the proper execution of a job or workflow. One way of achieving this, implicit in many trusted computing proposals, is to use application whitelisting to prevent unknown and untrusted applications from being executed on remote services. However, this approach has significant shortcomings across multiple administrative domains, as conflicts between locally managed whitelists will result in many useful services appearing untrustworthy to users. This has the potential to limit availability and prevent trusted distributed systems from ever being successfully deployed. We propose a set of requirements for a system which will manage these conflicts, and provide a mechanism for updating application whitelists that will increase service availability and trustworthiness. We also suggest and specify a set of components, including a centralised configuration manager, which will meet these requirements.

Trustworthy Distributed Systems Through Integrity-Reporting

With the growing influence of e-Science, substantial quantities of research are being facilitated, recorded, and reported by means of distributed computing. As a result, the scope for malicious intervention continues to grow and so do the rewards available to those able to steal the models and data that have significant commercial value. Researchers are often reluctant to exploit the full benefits of distributed computing because they fear the compromise of their sensitive data or the uncertainty of the returned results. In this chapter, we propose two types of trustworthy distributed systems – one suitable for a computational system and the other for a distributed data system. Central to these systems is the novel idea of configuration resolver, which, in both designs, is responsible for filtering trustworthy hosts and ensuring that jobs are dispatched to those considered trustworthy. Furthermore, the blind analysis server enables statistical analyses to be performed on sensitive raw data – collected from multiple sites – without disclosing it to anyone.