Abraham Matta

is a future Internet architecture based on the repeated use of a single, flexible functional unit for different capabilities over different scopes of a communication service. DRUID allows common protocol functions and capabilities to be reused from within a single block, avoiding the need for recapitulated implementation, and allowing these func-tions and scopes to be dynamically determined, enabling the service to adapt to changes in the local machine and

All truth passes through three stages. First, it is ridiculed. Second, it is violently opposed. Third, it is accepted as being self-evident. Arthur Schopenhauer iii Acknowledgments First and foremost I would like to thank my advisor, professor Abraham Matta, for his support, guidance and infinite patience. For that I am truly indebted. I would also like to thank professor Azer Bestavros, Ashwin Sridharan and Hui Zang for their very useful feedback and input. Finally, I would like to thank my parents and my brother for always being there for me. iv

To my friends, that bring light to my life even in its darkest hours. iv Acknowledgments The PhD journey is different for everyone but always full of experiences and lessons. The biggest lesson for me was that I can not do everything alone. This would not have been a successful journey if it has not been for the help of many different people that gave me the support I needed in order to finish what I started many years ago. First of all I would like to thank my advisor Abraham Matta, for his patience and guidance throughout the years. He gave me the freedom to find my own path and was always there to help and advise me. I would also like to thank my co-advisor Craig Partridge that taught me how to properly document and found my work. He also showed me how to leverage prior research to expand my knowledge and understanding of a research field. Both of my advisors are great teachers and I was very fortunate to work with them. The rest of my committee – Azer, Jason, John and Mark – als...

Quality control You are reviewing to accept, not to reject! You recommend for or against Useless without justification Accept if novel, relevant, validated, well written, worthy You are helping the final decision maker(s) Useless without evidence you understood the work Be honest, e.g. if you haven’t checked a part of the paper You are helping the authors; your colleagues Give suggestions for improvement Be constructive – Do unto others as you would have them do unto you

The Internet consists of thousands of autonomous systems (A Ses). Each AS represents an Internet Service Provider or the network of a large organization, that is managed independently. The Border Gateway Protocol is the policy routing protocol of ch oice for connecting these ASes, while allowing them to set their routing policies independently. Routing policies (or path preferences) determine how a path to a particular destination is chosen ou t of a candidate set of paths. This flexibility in configuring routing policies comes at the cost f stability, where ASes may have conflicting policies causing them to continually advertise new routing updates for extended periods of time. We introduce a theoretical framework for policy routing dyn amics (i.e., how path changes propagate in the network) that is based on the specifics of rou ting pdate mechanisms. Unlike existing models, our Dynamic Policy Routing (DPR) model intro duces several structures that capture how path changes propagat...

Commonly, research work in routing for delay tolerant networks (DTN) assumes that node encounters are predestined, in the sense that they are the result of unknown, exogenous processes that control the mobility of these nodes. In this paper, we argue that for many applications such an assumption is too restrictive: while the spatio-temporal coordinates of the start and end points of a node’s journey are determined by exogenous processes, the specific path that a node may take in space-time, and hence the set of nodes it may encounter could be controlled in such a way so as to improve the performance of DTN routing. To that end, we consider a setting in which each mobile node is governed by a schedule consisting of a list of locations that the node must visit at particular times. Typically, such schedules exhibit some level of slack, which could be leveraged for DTN message delivery purposes. We define the Mobility Coordination Problem (MCP) for DTNs as follows: Given a set of nodes,...

With the increased use of “Virtual Machines ” (VMs) as vehicles that isolate applications running on the same host, it is necessary to devise techniques that enable multiple VMs to share underlying resources both fairly and efficiently.To that end, one common approach is to deploy complex resource management techniques in the hosting infrastructure.Alternately, in this paper, we advocate the use of selfadaptation in the VMs themselves based on feedback about resource usage and availability.Consequently, we define a “Friendly ” VM (FVM) to be a virtual machine that adjusts its demand for system resources, so that they are both efficiently and fairly allocated to competing FVMs.Such properties are ensured using one of many provably convergent control rules, such as Additive-Increase/Multiplicative-Decrease (AIMD).By adopting this distributed applicationbased approach to resource management, it is not necessary to make assumptions about the underlying resources nor about the requiremen...

Network virtualization has enabled new business models by allowing infrastructure providers to lease or share their physical network. To concurrently run multiple customized virtual network services, such infrastructure providers need to run a virtual network embedding protocol. The virtual network embedding is the (NP-hard) problem of matching constrained virtual networks onto the physical network. We present the design and implementation of the first policy-based architecture for the virtual network embedding problem. By policy, we mean a variant aspect of any of the (invariant) embedding mechanisms: resource discovery, virtual network mapping, and allocation on the physical infrastructure. Our architecture adapts to different scenarios by instantiating appropriate policies, and has bounds on embedding efficiency and on convergence embedding time, over a single provider, or across multiple federated providers. The performance of representative novel policy configurations are compa...

We revisit the problem of connection management for reliable transport as part of our clean-slate Recursive InterNet Architecture (RINA) [5]. At one extreme, a pure soft-state (SS) approach (as in Delta-t [15]) safely removes the state of a connection at the sender and receiver once the state timers expire without the need for explicit removal messages. And new connections are established without an explicit handshaking phase. On the other hand, a hybrid hard-state/soft-state (HS+ SS) approach (as in TCP) uses both explicit handshaking as well as more limited timer-based management of the connection’s state. In this paper, we consider the worst-case scenario of reliable single-message communication. Using simulation, we evaluate various approaches in terms of correctness (with respect to dat loss and duplication) and robustness to bad network conditions (high message loss rate and variable channel delays). Our results show that the SS approach is more robust, and has lower message o...

Page 1. 2007 15th IEEE Workshop on Local and Metropolitan Area Networks Referees Victor Bahl Maria Papadopouli Andrea Bianco Krishna Pattabhiraman Ernst Biersack Daniel Pitt Flavio Bonomi George Polyzos Jack Brassil Lili Qiu Torsten Braun KK Ramakrishnan Reuven Cohen Byrav Ramamurthy Virgil Dobrota Gopalakrishnan Ramamurthy Robert Doverspike Kaliappa Ravindran Aveek Dutta George Rouskas Stein Gjessing Puneet Sharma Aloke Guha Vishal Sharma Bjarne Hellvik Vasilios Siris Enrique Hernandez-Valencia Ioannis Sourdis ...

When analysing the behavior of complex networked systems, it is often the case that some components within that network are only known to the extent that they belong to one of a set of possible" implementations"–eg, versions of a specific protocol, class of schedulers, etc. In this report we augment the specification language considered in BUCSTR-2004-021, BUCS-TR-2005-014, BUCS-TR-2005-015, and BUCS-TR-2005-033, to include a non-deterministic multiple-choice let-binding, which allows us to consider compositions of networking ...

The best-effort nature of the Internet poses a significant obstacle to the deployment of many applications that require guaranteed bandwidth. We present a novel approach that enables two edge/border routers which we call Internet traffic managers (ITM)-to use an adaptive number of TCP connections to set up a tunnel of desirable bandwidth between them. The number of TCP connections that comprise this tunnel is elastic in the sense that it increases/decreases in tandem with competing cross traffic to maintain a target bandwidth ...

We present a type inference algorithm, in the style of compositional analysis, for the language TRAFFIC—a specification language for flow composition applications proposed in [2]—and prove that this algorithm is correct: the typings it infers are principal typings, and the typings agree with syntax-directed type checking on closed flow specifications. This algorithm is capable of verifying partial flow specifications, which is a significant improvement over syntax-directed type checking algorithm presented in [3]. We also show ...