Gowtham Bellala

Application-awareness in SDN

We present a framework, Atlas, which incorporates application-awareness into Software-Defined Networking (SDN), which is currently capable of L2/3/4-based policy enforcement but agnostic to higher layers. Atlas enables fine-grained, accurate and scalable application classification in SDN. It employs a machine learning (ML) based traffic classification technique, a crowd-sourcing approach to obtain ground truth data and leverages SDNs data reporting mechanism and centralized control. We prototype Atlas on HP Labs wireless networks and observe 94% accuracy on average, for top 40 Android applications.

Following the electrons: methods for power management in commercial buildings

Commercial buildings are significant consumers of electricity. The first step towards better energy management in commercial buildings is monitoring consumption. However, instrumenting every electrical panel in a large commercial building is expensive and wasteful. In this paper, we propose a greedy meter (sensor) placement algorithm based on maximization of information gained, subject to a cost constraint. The algorithm provides a near-optimal solution guarantee. Furthermore, to identify power saving opportunities, we use an unsupervised anomaly detection technique based on a low-dimensional embedding. Further, to better manage resources such as lighting and HVAC, we propose a semi-supervised approach combining hidden Markov models (HMM) and a standard classifier to model occupancy based on readily available port-level network statistics.

A Rank-Based Approach to Active Diagnosis

The problem of active diagnosis arises in several applications such as disease diagnosis and fault diagnosis in computer networks, where the goal is to rapidly identify the binary states of a set of objects (e.g., faulty or working) by sequentially selecting, and observing, potentially noisy responses to binary valued queries. Previous work in this area chooses queries sequentially based on Information gain, and the object states are inferred by maximum a posteriori (MAP) estimation. In this work, rather than MAP estimation, we aim to rank objects according to their posterior fault probability. We propose a greedy algorithm to choose queries sequentially by maximizing the area under the ROC curve associated with the ranked list. The proposed algorithm overcomes limitations of existing work. When multiple faults may be present, the proposed algorithm does not rely on belief propagation, making it feasible for large scale networks with little loss in performance. When a single fault is present, the proposed algorithm can be implemented without knowledge of the underlying query noise distribution, making it robust to any misspecification of these noise parameters. We demonstrate the performance of the proposed algorithm through experiments on computer networks, a toxic chemical database, and synthetic datasets.

A finite state machine-based characterization of building entities for monitoring and control

Cyber physical systems such as buildings contain entities (devices, appliances, etc.) that consume a multitude of resources (power, water, etc.). Efficient operation of these entities is important for reducing operating costs and environmental footprint of buildings. In this paper, we propose an entity characterization framework based on a finite state machine abstraction. Each state in the state machine is characterized in terms of distributions of sustainability or performance metrics of interest. This framework provides a basis for anomaly detection, assessment, prediction and usage pattern discovery. We demonstrate the usefulness of the framework using data from actual building entities. In particular, we apply our methodology to chillers and cooling towers, components of a building HVAC system.

Unlocking proteomic heterogeneity in complex diseases through visual analytics

Despite years of preclinical development, biological interventions designed to treat complex diseases such as asthma often fail in phase III clinical trials. These failures suggest that current methods to analyze biomedical data might be missing critical aspects of biological complexity such as the assumption that cases and controls come from homogeneous distributions. Here we discuss why and how methods from the rapidly evolving field of visual analytics can help translational teams (consisting of biologists, clinicians, and bioinformaticians) to address the challenge of modeling and inferring heterogeneity in the proteomic and phenotypic profiles of patients with complex diseases. Because a primary goal of visual analytics is to amplify the cognitive capacities of humans for detecting patterns in complex data, we begin with an overview of the cognitive foundations for the field of visual analytics. Next, we organize the primary ways in which a specific form of visual analytics called networks has been used to model and infer biological mechanisms, which help to identify the properties of networks that are particularly useful for the discovery and analysis of proteomic heterogeneity in complex diseases. We describe one such approach called subject‐protein networks, and demonstrate its application on two proteomic datasets. This demonstration provides insights to help translational teams overcome theoretical, practical, and pedagogical hurdles for the widespread use of subject‐protein networks for analyzing molecular heterogeneities, with the translational goal of designing biomarker‐based clinical trials, and accelerating the development of personalized approaches to medicine.