Professor of Computer ScienceI dedicate this thesis to my beautiful daughter Gigi, and my lovely ... more Professor of Computer ScienceI dedicate this thesis to my beautiful daughter Gigi, and my lovely wife Houry who supported me each step of the way. A special dedication goes to my father Raffi, my sister Mary, and my late mother Georgette whose memory fills me with strength and drives me forward. iv Acknowledgments A PhD journey is a long, rewarding, but full of frustrating moments. I would not be able to complete it without the love and support of my wife, and the encouragement of my family. I would like to thank my advisor Professor Azer Bestavros for his guidance and advice. He is truely a great role model. I would also like to thank my academic advisor Professor Assaf kfoury for guiding me during my early PhD years. I was fortunate to have faculty collaborators at Boston University whom I owe a debt of gradtitude. I would like to thank professors Abrahim Matta, Evimaria Terzi, Ben Lubin, Jonathan Appavoo, and Orran Krieger for serving on my committee and for all their positive fe...
Background and Motivation" section to specify that we were discussing fi-SN proofs for type ... more Background and Motivation" section to specify that we were discussing fi-SN proofs for type systems more powerful than the simply-typed -calculus, the wording in the rest of the paper was not so careful and a comment which was intended to go at the beginning of Section 5 was placed at the beginning of Section 4 instead. As a result, the paper erroneously made false claims about the history of proving the fi-SN property for the simply-typed -calculus. Many readers have pointed out numerous methods for proving the fi-SN property for the simply-typed -calculus that are not variations on the methods of Tait or Girard. The methods mentioned have included proofs by Gandy, Nederpelt, Klop, de Groote, de Vrijer, van de Pol and Schwichtenberg [Gan80b, Ned73, Klo80, dG93, dV87, vdPS95]. Our intention was to discuss proofs of the fi-SN property for more powerful type systems such as system F,
Openflow provides a standard interface for separating a network into a data plane and a programma... more Openflow provides a standard interface for separating a network into a data plane and a programmatic control plane. This enables easy network reconfiguration, but introduces the potential for programming bugs to cause network effects. To study OpenFlow switch behavior, we used Alloy to create a software abstraction describing the internal state of a network and its OpenFlow switches. This work is an attempt to model the static and dynamic behaviour a network built using OpenFlow switches.
A reassembling of a simple graph G = (V,E) is an abstraction of a problem arising in earlier stud... more A reassembling of a simple graph G = (V,E) is an abstraction of a problem arising in earlier studies of network analysis. There are several equivalent definitions of graph reassembling; in this report we use a definition which makes it closest to the notion of graph carving. A reassembling is a rooted binary tree whose nodes are subsets of V and whose leaf nodes are singleton sets, with each of the latter containing a distinct vertex of G. The parent of two nodes in the reassembling is the union of the two children's vertex sets. The root node of the reassembling is the full set V. The edge-boundary degree of a node in the reassembling is the number of edges in G that connect vertices in the node's set to vertices not in the node's set. A reassembling's alpha-measure is the largest edge-boundary degree of any node in the reassembling. A reassembling of G is alpha-optimal if its alpha-measure is the minimum among all alpha-measures of G's reassemblings. The proble...
We present elements of a typing theory for flow networks, where "types", "typings&... more We present elements of a typing theory for flow networks, where "types", "typings", and "type inference" are formulated in terms of familiar notions from polyhedral analysis and convex optimization. Based on this typing theory, we develop an alternative approach to the design and analysis of network algorithms, which we illustrate by applying it to the max-flow problem in multiple-source, multiple-sink, capacited directed planar graphs.
Formal correctness of complex multi-party protocols can be difficult to verify. While models of s... more Formal correctness of complex multi-party protocols can be difficult to verify. While models of specific sign constraints, protocols which lend themselves to arbitrarily many compositions of agents-such as the chaining of proxies or the peering of routers-are more difficult to verify because they represent potentially infinite state spaces and may exhibit emergent behaviors which may not materialize under particular fixed compositions. We address this challenge by developing an algebraic approach that enables us to reduce ...
We pull together previously established graph-theoretical results to produce the algorithm in the... more We pull together previously established graph-theoretical results to produce the algorithm in the paper's title. The glue are three easy elementary lemmas.
We prove that first order logic is strictly weaker than fixed point logic over every infinite cla... more We prove that first order logic is strictly weaker than fixed point logic over every infinite classes of finite ordered structures with additional unary relations: Over these classes there is always an inductive unary relation which cannot be defined by a first-order formula, even when every inductive sentence (i.e., closed formula) can be expressed in first-order over this particular class. Our proof first establishes a property valid for every unary relation definable by first-order logic over these classes which is peculiar to classes of ordered structures with unary relations. In a second step we show that this property itself can be expressed in fixed point logic and can be used to construct a non-elementary unary relation.
We envision future Sensor Networks (SNs) that will be composed of a hybrid collection of a variet... more We envision future Sensor Networks (SNs) that will be composed of a hybrid collection of a variety of sensing devices embedded into shared environments. In such environments it follows that the embedded SN infrastructure would also be shared by various users, occupants, or administrators of these shared spaces. As such a clear need emerges to virtualize the SN, sharing the resources of the SN across various tasks executing simultaneously. To achieve this goal, we present the sn B ench (SN Workbench). The sn B ...
We envision the emergence of general-purpose, well-provisioned sensor networks-which we call"... more We envision the emergence of general-purpose, well-provisioned sensor networks-which we call" Sensoria"-that are embedded in (or overlayed atop) physical spaces, and whose use is shared amongst autonomous users of that space for independent and possibly conflicting missions. Our conception of a Sensorium stands in sharp contrast to the commonly adopted view of an embedded sensor network as a special-purpose infrastructure that serves a well-defined, fixed mission. The usefulness of a Sensorium will ...
Professor of Computer ScienceI dedicate this thesis to my beautiful daughter Gigi, and my lovely ... more Professor of Computer ScienceI dedicate this thesis to my beautiful daughter Gigi, and my lovely wife Houry who supported me each step of the way. A special dedication goes to my father Raffi, my sister Mary, and my late mother Georgette whose memory fills me with strength and drives me forward. iv Acknowledgments A PhD journey is a long, rewarding, but full of frustrating moments. I would not be able to complete it without the love and support of my wife, and the encouragement of my family. I would like to thank my advisor Professor Azer Bestavros for his guidance and advice. He is truely a great role model. I would also like to thank my academic advisor Professor Assaf kfoury for guiding me during my early PhD years. I was fortunate to have faculty collaborators at Boston University whom I owe a debt of gradtitude. I would like to thank professors Abrahim Matta, Evimaria Terzi, Ben Lubin, Jonathan Appavoo, and Orran Krieger for serving on my committee and for all their positive fe...
Background and Motivation" section to specify that we were discussing fi-SN proofs for type ... more Background and Motivation" section to specify that we were discussing fi-SN proofs for type systems more powerful than the simply-typed -calculus, the wording in the rest of the paper was not so careful and a comment which was intended to go at the beginning of Section 5 was placed at the beginning of Section 4 instead. As a result, the paper erroneously made false claims about the history of proving the fi-SN property for the simply-typed -calculus. Many readers have pointed out numerous methods for proving the fi-SN property for the simply-typed -calculus that are not variations on the methods of Tait or Girard. The methods mentioned have included proofs by Gandy, Nederpelt, Klop, de Groote, de Vrijer, van de Pol and Schwichtenberg [Gan80b, Ned73, Klo80, dG93, dV87, vdPS95]. Our intention was to discuss proofs of the fi-SN property for more powerful type systems such as system F,
Openflow provides a standard interface for separating a network into a data plane and a programma... more Openflow provides a standard interface for separating a network into a data plane and a programmatic control plane. This enables easy network reconfiguration, but introduces the potential for programming bugs to cause network effects. To study OpenFlow switch behavior, we used Alloy to create a software abstraction describing the internal state of a network and its OpenFlow switches. This work is an attempt to model the static and dynamic behaviour a network built using OpenFlow switches.
A reassembling of a simple graph G = (V,E) is an abstraction of a problem arising in earlier stud... more A reassembling of a simple graph G = (V,E) is an abstraction of a problem arising in earlier studies of network analysis. There are several equivalent definitions of graph reassembling; in this report we use a definition which makes it closest to the notion of graph carving. A reassembling is a rooted binary tree whose nodes are subsets of V and whose leaf nodes are singleton sets, with each of the latter containing a distinct vertex of G. The parent of two nodes in the reassembling is the union of the two children's vertex sets. The root node of the reassembling is the full set V. The edge-boundary degree of a node in the reassembling is the number of edges in G that connect vertices in the node's set to vertices not in the node's set. A reassembling's alpha-measure is the largest edge-boundary degree of any node in the reassembling. A reassembling of G is alpha-optimal if its alpha-measure is the minimum among all alpha-measures of G's reassemblings. The proble...
We present elements of a typing theory for flow networks, where "types", "typings&... more We present elements of a typing theory for flow networks, where "types", "typings", and "type inference" are formulated in terms of familiar notions from polyhedral analysis and convex optimization. Based on this typing theory, we develop an alternative approach to the design and analysis of network algorithms, which we illustrate by applying it to the max-flow problem in multiple-source, multiple-sink, capacited directed planar graphs.
Formal correctness of complex multi-party protocols can be difficult to verify. While models of s... more Formal correctness of complex multi-party protocols can be difficult to verify. While models of specific sign constraints, protocols which lend themselves to arbitrarily many compositions of agents-such as the chaining of proxies or the peering of routers-are more difficult to verify because they represent potentially infinite state spaces and may exhibit emergent behaviors which may not materialize under particular fixed compositions. We address this challenge by developing an algebraic approach that enables us to reduce ...
We pull together previously established graph-theoretical results to produce the algorithm in the... more We pull together previously established graph-theoretical results to produce the algorithm in the paper's title. The glue are three easy elementary lemmas.
We prove that first order logic is strictly weaker than fixed point logic over every infinite cla... more We prove that first order logic is strictly weaker than fixed point logic over every infinite classes of finite ordered structures with additional unary relations: Over these classes there is always an inductive unary relation which cannot be defined by a first-order formula, even when every inductive sentence (i.e., closed formula) can be expressed in first-order over this particular class. Our proof first establishes a property valid for every unary relation definable by first-order logic over these classes which is peculiar to classes of ordered structures with unary relations. In a second step we show that this property itself can be expressed in fixed point logic and can be used to construct a non-elementary unary relation.
We envision future Sensor Networks (SNs) that will be composed of a hybrid collection of a variet... more We envision future Sensor Networks (SNs) that will be composed of a hybrid collection of a variety of sensing devices embedded into shared environments. In such environments it follows that the embedded SN infrastructure would also be shared by various users, occupants, or administrators of these shared spaces. As such a clear need emerges to virtualize the SN, sharing the resources of the SN across various tasks executing simultaneously. To achieve this goal, we present the sn B ench (SN Workbench). The sn B ...
We envision the emergence of general-purpose, well-provisioned sensor networks-which we call"... more We envision the emergence of general-purpose, well-provisioned sensor networks-which we call" Sensoria"-that are embedded in (or overlayed atop) physical spaces, and whose use is shared amongst autonomous users of that space for independent and possibly conflicting missions. Our conception of a Sensorium stands in sharp contrast to the commonly adopted view of an embedded sensor network as a special-purpose infrastructure that serves a well-defined, fixed mission. The usefulness of a Sensorium will ...
Uploads