Marco Di Bartolomeo

Dynamic Traceroute Visualization at Multiple Abstraction Levels

We present a system, called TPlay, for the visualization of the traceroutes performed by the Internet probes deployed by active measurement projects. These traceroutes are continuously executed towards selected Internet targets. TPlay allows to look at traceroutes at different abstraction levels and to animate the evolution of traceroutes during a selected time interval. The system has been extensively tested on traceroutes performed by RIPE Atlas [22] Internet probes.

Anchored Drawings of Planar Graphs

In this paper we study the Anchored Graph Drawing AGD problem: Given a planar graph G, an initial placement for its vertices, and a distance d, produce a planar straight-line drawing of G such that each vertex is at distance at most d from its original position. We show that the AGD problem is NP-hard in several settings and provide a polynomial-time algorithm when d is the uniform distance L ∞ and edges are required to be drawn as horizontal or vertical segments.

L-Drawings of Directed Graphs

We introduce L-drawings, a novel paradigm for representing directed graphs aiming at combining the readability features of orthogonal drawings with the expressive power of matrix representations. In an L-drawing, vertices have exclusive x- and y-coordinates and edges consist of two segments, one exiting the source vertically and one entering the destination horizontally. We study the problem of computing L-drawings using minimum ink. We prove its NP-completeness and provide a heuristic based on a polynomial-time algorithm that adds a vertex to a drawing using the minimum additional ink. We performed an experimental analysis of the heuristic which confirms its effectiveness.

Discovering high-impact routing events using traceroutes

With the increasing diffusion of Internet probing technologies, a large amount of regularly collected traceroutes are available for Internet Service Providers (ISPs) at low cost. We introduce a practically applicable methodology and algorithm that, given solely an arbitrary set of traceroutes, spot routing paths that change similarly over time, aggregate them into inferred events, and report each event along with the impacted observation points and a small set of IP addresses that can help identify its cause. The formal model at the basis of our methodology revolves around the notion of empathy, a relation that binds similarly behaving traceroutes. The correctness and completeness of our approach are based on structural properties that are easily expressed in terms of empathic measurements. We perform experiments with data from public measurement infrastructures like RIPE Atlas, showing the effectiveness of our algorithm in distilling events from a large amount of traceroute data. We also validate the accuracy of the inferred events against ground-truth knowledge of routing changes originating from induced and spontaneous routing events. Given these promising results, we believe our methodology can be an effective aid for ISPs to detect and track routing changes affecting many users (with potentially adverse effects on their connection quality).

Is it really worth to peer at IXPs? A comparative study

Internet Exchange Points (IXPs) play a crucial role in the Internet ecosystem. However, existing literature fails in quantitatively assessing the advantage for an Internet Service Provider (ISP) to peer at an IXP. We give a contribution to bridge such a gap by collaborating with three medium-sized ISPs in Italy to compare key performance indicators (round-trip delay, hop count, packet loss, and jitter) as measured from several vantage points in presence and absence of IXP peerings. Our findings are that IXP-based paths exhibit better and more stable performance, whereas avoiding IXPs introduces performance deterioration and higher variability. Moreover, our measurements confirm that IXP-based paths tend to preserve the locality of traffic.

Drawing Georeferenced Graphs - Combining Graph Drawing and Geographic Data

We consider the task of visually exploring relationships (such as established connections, similarity, reachability, etc) among a set of georeferenced entities, i.e., entities that have geographic data associated with them. A novel 2.5D paradigm is proposed that provides a robust and practical solution based on separating and then integrating back again the networked and geographical dimensions of the input dataset. This allows us to easily cope with partial or incomplete geographic annotations, to reduce cluttering of close entities, and to address focus-plus-context visualization issues. Typical application domains include, for example, coordinating search and rescue teams or medical evacuation squads, monitoring ad-hoc networks, exploring location-based social networks and, more in general, visualizing relational datasets including geographic annotations.

Radian: Visual Exploration of Traceroutes

Several projects deploy probes in the Internet. Probes are systems that continuously perform traceroutes and other networking measurements (e.g., ping) towards selected targets. Measurements can be stored and analyzed to gain knowledge on several aspects of the Internet, but making sense of such data requires suitable methods and tools for exploration and visualization. We present Radian, a tool that allows to visualize traceroute paths at different levels of detail and to animate their evolution during a selected time interval. We also describe extensive tests of the tool using traceroutes performed by RIPE Atlas Internet probes.

Cheapest paths in dynamic networks

Flows over time problems relate to finding optimal flows over a capacitated network where transit times on network arcs are explicitly considered. In this article, we study the problem of determining a minimum cost origin-destination path where the cost and the travel time of each arc depend on the time taken to travel from the origin to that particular arc along the path. We provide computational complexity results for this problem and an exact solution algorithm based on an enumeration scheme on the corresponding time expanded network. Finally, we show the efficiency of our approach through a number of experimental tests.

RoutingWatch: Visual exploration and analysis of routing events

Network operators invest significant resources in monitoring and troubleshooting the infrastructure they run. Currently, the availability of large networks of probing devices, like, for example, RIPE Atlas, dramatically increases the amount of data an operator can rely on. In particular, the large amount of traceroutes they can produce are both an opportunity and a challenge. In this paper we provide a detailed description of RoutingWatch, a visual tool for interactively performing searches and analysis of routing events inferred from a large set of traceroutes. The key design choices of RoutingWatch are based on discussions with network operators. We evaluate the effectiveness of our approach by conducting a preliminary user study.

There is more to streamgraphs than movies: better aesthetics via ordering and lassoing

Streamgraphs were popularized in 2008 when The New York Times used them to visualize box office revenues for 7500 movies over 21 years. The aesthetics of a streamgraph is affected by three components: the ordering of the layers, the shape of the lowest curve of the drawing, known as the baseline, and the labels for the layers. As of today, the ordering and baseline computation algorithms proposed in the paper of Byron and Wattenberg are still considered the state of the art. However, their ordering algorithm exploits statistical properties of the movie revenue data that may not hold in other data . In addition, the baseline optimization is based on a definition of visual energy that in some cases results in considerable amount of visual distortion. We offer an ordering algorithm that works well regardless of the properties of the input data , and propose a 1‐norm based definition of visual energy and the associated solution method that overcomes the limitation of the original baseline optimization procedure. Furthermore, we propose an efficient layer labeling algorithm that scales linearly to the data size in place of the brute‐force algorithm adopted by Byron and Wattenberg. We demonstrate the advantage of our algorithms over existing techniques on a number of real world data sets.