Claudio Martella

How Well Do Graph-Processing Platforms Perform? An Empirical Performance Evaluation and Analysis

Graph-processing platforms are increasingly used in a variety of domains. Although both industry and academia are developing and tuning graph-processing algorithms and platforms, the performance of graph-processing platforms has never been explored or compared in-depth. Thus, users face the daunting challenge of selecting an appropriate platform for their specific application. To alleviate this challenge, we propose an empirical method for benchmarking graph-processing platforms. We define a comprehensive process, and a selection of representative metrics, datasets, and algorithmic classes. We implement a benchmarking suite of five classes of algorithms and seven diverse graphs. Our suite reports on basic (user-lever) performance, resource utilization, scalability, and various overhead. We use our benchmarking suite to analyze and compare six platforms. We gain valuable insights for each platform and present the first comprehensive comparison of graph-processing platforms.

Adaptive partitioning for large-scale dynamic graphs

In the last years, large-scale graph processing has gained increasing attention, with most recent systems placing particular emphasis on latency. One possible technique to improve runtime performance in a distributed graph processing system is to reduce network communication. The most notable way to achieve this goal is to partition the graph by minimizing the number of edges that connect vertices assigned to different machines, while keeping the load balanced. However, real-world graphs are highly dynamic, with vertices and edges being constantly added and removed. Carefully updating the partitioning of the graph to reflect these changes is necessary to avoid the introduction of an extensive number of cut edges, which would gradually worsen computation performance. In this paper we show that performance degradation in dynamic graph processing systems can be avoided by adapting continuously the graph partitions as the graph changes. We present a novel highly scalable adaptive partitioning strategy, and show a number of refinements that make it work under the constraints of a large-scale distributed system. The partitioning strategy is based on iterative vertex migrations, relying only on local information. We have implemented the technique in a graph processing system, and we show through three real-world scenarios how adapting graph partitioning reduces execution time by over 50% when compared to commonly used hash-partitioning.

Benchmarking graph-processing platforms: a vision

Processing graphs, especially at large scale, is an increasingly useful activity in a variety of business, engineering, and scientific domains. Already, there are tens of graph-processing platforms, such as Hadoop, Giraph, GraphLab, etc., each with a different design and functionality. For graph-processing to continue to evolve, users have to find it easy to select a graph-processing platform, and developers and system integrators have to find it easy to quantify the performance and other non-functional aspects of interest. However, the state of performance analysis of graph-processing platforms is still immature: there are few studies and, for the few that exist, there are few similarities, and relatively little understanding of the impact of dataset and algorithm diversity on performance. Our vision is to develop, with the help of the performance-savvy community, a comprehensive benchmarking suite for graph-processing platforms. In this work, we take a step in this direction, by proposing a set of seven challenges, summarizing our previous work on performance evaluation of distributed graph-processing platforms, and introducing our on-going work within the SPEC Research Groups Cloud Working Group.

Crowd textures as proximity graphs

We are only starting to understand how people behave when they are part of a crowd. This article presents a novel approach to the study and management of crowds. The approach comprises a device to be worn by individuals, an infrastructure to collect the information from the devices, a set of algorithms for recognizing crowd dynamics, and a set of feedback strategies to intervene in the crowd. A fundamental element of our approach is to consider crowds in terms of their texture. The crowd texture is represented through the proximity graph, a data structure that captures the spatial closeness relationship between individuals over time. We address its properties and limitations, a system architecture to measure and process it, and a few examples of insights that can be obtained from analyzing it.

From proximity sensing to spatio-temporal social graphs

Understanding the social dynamics of a group of people can give new insights into social behavior. Physical proximity between individuals results from the interactions between them. Hence, measuring physical proximity is an important step towards a better understanding of social behavior. We discuss a novel approach to sense proximity from within the social dynamics. Our primary objective is to construct a spatio-temporal social graph from noisy proximity data. We address the technical and algorithmic challenges of measuring proximity reliably and accurately. Simulations and real world experiments demonstrate the feasibility and scalability of our approach. Our algorithms doubles the sensitivity of proximity detections at the cost of a slight reduction in specificity.

Leveraging proximity sensing to mine the behavior of museum visitors

Face-to-face proximity has been successfully leveraged to study the relationships between individuals in various contexts, from a working place, to a conference, a museum, a fair, and a date. We spend time facing the individuals with whom we chat, discuss, work, and play. However, face-to-face proximity is not the realm of solely person-to-person relationships, but it can be used as a proxy to study person-to-object relationships as well. We face the objects with which we interact on a daily basis, like a television, the kitchen appliances, a book, including more complex objects like a stage where a concert is taking place. In this paper, we focus on the relationship between the visitors of an art exhibition and its exhibits. We design, implement, and deploy a sensing infrastructure based on inexpensive mobile proximity sensors and a filtering pipeline that we use to measure face-to-face proximity between individuals and exhibits. Our pipeline produces an improvement in measurement accuracy of up to 64% relative to raw data. We use this data to mine the behavior of the visitors and show that group behavior can be recognized by means of data clustering and visualization.

How Was It?: Exploiting Smartphone Sensing to Measure Implicit Audience Responses to Live Performances

In this paper, we present an approach to understand the response of an audience to a live dance performance by the processing of mobile sensor data. We argue that exploiting sensing capabilities already available in smart phones enables a potentially large scale measurement of an audiences implicit response to a performance. In this work, we leverage both tri-axial accelerometers, worn by ordinary members of the public during a dance performance, to predict responses to a number of survey answers, comprising enjoyment, immersion, willingness to recommend the event to others, and change in mood. We also analyse how behaviour as a result of seeing a dance performance might be reflected in a peoples subsequent social behaviour using proximity and acceleration sensing. To our knowledge, this is the first work where pervasive mobile sensing has been used to investigate spontaneous responses to predict the affective evaluation of a live performance. Using a single body worn accelerometer to monitor a set of audience members, we were able to predict whether they enjoyed the event with a balanced classification accuracy of 90\%. The collective coordination of the audiences bodily movements also highlighted memorable moments that were reported later by the audience. The effective use of body movements to measure affective responses in such a setting is particularly surprising given that traditionally, physiological signals such as skin conductance or brain-based signals are the more commonly accepted methods to measure implicit affective response. Our experiments open interesting new directions for research on both automated techniques and applications for the implicit tagging of real world events via spontaneous and implicit audience responses during as well as after a performance.

Visualizing, clustering, and predicting the behavior of museum visitors ☆

Fine-arts museums design exhibitions to educate, inform and entertain visitors. Existing work leverages technology to engage, guide and interact with the visitors, neglecting the need of museum staff to understand the response of the visitors. Surveys and expensive observational studies are currently the only available data source to evaluate visitor behavior, with limits of scale and bias. In this paper, we explore the use of data provided by low-cost mobile and fixed proximity sensors to understand the behavior of museum visitors. We present visualizations of visitor behavior, and apply both clustering and prediction techniques to the collected data to show that group behavior can be identified and leveraged to support the work of museum staff.

Exploiting Density to Track Human Behavior in Crowded Environments

For the Internet of Things to be people-centered, things need to identify when people and their things are nearby. In this article, we present the design, implementation, and deployment of a positioning system based on mobile and fixed inexpensive proximity sensors that we use to track when individuals are close to an instrumented object or placed at certain points of interest. To overcome loss of data between mobile and fixed sensors due to crowd density, traditional approaches are extended with mobile-to-mobile proximity information. We tested our system in a museum crowded with thousands of visitors, showing that measurement accuracy increases in the presence of more individuals wearing a proximity sensor. Furthermore, we show that density information can be leveraged to study the behavior of the visitors, for example, to track the popularity of points of interest, and the flow and distribution of visitors across floors.

Working with Giraph

Previous chapters introduced the generic Giraph programming model and talked about a few common use cases that lend themselves easily to being modeled as graph-processing applications. This chapter covers practical aspects of developing Giraph applications and focuses on running Giraph on top of Hadoop.