Avinash Vyas

Enabling context-aware and privacy-conscious user data sharing

This paper provides detail on two key components of the Houdini framework under development at Bell Labs, that enable context-aware and privacy-conscious user data sharing appropriate for mobile and/or ubiquitous computing. The framework includes an approach for integrating data from diverse sources, for gathering user preferences for what data to share and when to share it, and a policy management infrastructure in the network for enforcing those preferences. The current paper focuses on two components of this infrastructure that are essential for mobile and ubiquitous computing, namely the framework to support self-provisioning of preferences, and the performance of the underlying rules engine.

Verification of tree updates for optimization

With the rise of XML as a standard format for representing tree-shaped data, new programming tools have emerged for specifying transformations to tree-like structures. A recent example along this line are the update languages of [16,15,8] which add tree update primitives on top of the declarative query languages XPath and XQuery. These tree update languages use a “snapshot semantics”, in which all querying is performed first, after which a generated sequence of concrete updates is performed in a fixed order determined by query evaluation. In order to gain efficiency, one would prefer to perform updates as soon as they are generated, before further querying. This motivates a specific verification problem: given a tree update program, determine whether generated updates can be performed before all querying is completed. We formalize this notion, which we call “Binding Independence”. We give an algorithm to verify that a tree update program is Binding Independent, and show how this analysis can be used to produce optimized evaluation orderings that significantly reduce processing time.

Policy-aware sender anonymity in location based services

Sender anonymity in location-based services (LBS) attempts to hide the identity of a mobile device user who sends requests to the LBS provider for services in her proximity (e.g. “find the nearest gas station” etc.). The goal is to keep the requesters interests private even from attackers who (via hacking or subpoenas) gain access to the request and to the locations of the mobile user and other nearby users at the time of the request. In an LBS context, the best-studied privacy guarantee is known as sender k-anonymity. We show that state-of-the art solutions for sender k-anonymity defend only against naive attackers who have no knowledge of the anonymization policy that is in use. We strengthen the privacy guarantee to defend against more realistic “policy-aware” attackers. We describe a polynomial algorithm to obtain an optimum anonymization policy. Our implementation and experiments show that the policy-aware sender k-anonymity has potential for practical impact, being efficiently enforceable, with limited reduction in utility when compared to policy-unaware guarantees.

IMPROVING USER EXPERIENCE THROUGH RULE-BASED SERVICE CUSTOMIZATION

The web and converged services paradigm promises tremendous flexibility in the creation of rich composite services for enterprises and end users. The flexibility and richness offers the possibility of highly customized, individualized services for the end user and hence revenue generating services for service providers (e.g. ASPs, telecom network operators, ISPs). But how can end users (and enterprises) specify their preferences when a myriad of possibilities and potential circumstances need to be addressed? In this paper, we advocate a solution based on policy management where user preferences are specified through forms but translated into rules in a high-level policy language. This paper identifies the requirements for this kind of interpretation, and describes the Houdini system (developed at Bell Labs) which offers a rich rule-based language and a framework that supports intuitive, forms-based provisioning interfaces.

A multi-layer dynamic model for customer experience analytics

Today, data collected by service providers can track an individual users experience in detail, at flow or packet level in real time. However, we still lack analytics methods that can translate this information into a comprehensive and ever-evolving representation of the user experience. In this paper, we provide a layered dynamic model that addresses the problem of how to relate low-level network performance metrics to a users perception of network service and their subsequent actions. Using time-stamped observations from networks, devices, and customer care, we build probabilistic models to link network performance to an inferred state of customer satisfaction, and then to explicit and implicit customer disengagement events. We provide inference algorithms for the model parameters, and report test results on synthesized datasets based on real, but incomplete, observations. We discuss how popular anonymization techniques such as data masking, encryption, k-anonymization, and differential privacy can be used to protect sensitive and private user data without impacting the user experience inference.