arXiv subjects

Chen Avin

Publications and source records attributed to Chen Avin.

At least 37 records · Page 2Linked to original sources

arXiv2020

Dynamic Balanced Graph Partitioning

This paper initiates the study of the classic balanced graph partitioning problem from an online perspective: Given an arbitrary sequence of pairwise communication requests between $n$ nodes, with patterns that may change over time, the objective is to service these requests efficiently by partitioning the nodes into $\ell$ clusters, each of size $k$, such that frequently communicating nodes are located in the same cluster. The partitioning can be updated dynamically by migrating nodes between clusters. The goal is to devise online algorithms which jointly minimize the amount of inter-cluster communication and migration cost. The problem features interesting connections to other well-known online problems. For example, scenarios with $\ell=2$ generalize online paging, and scenarios with $k=2$ constitute a novel online variant of maximum matching. We present several lower bounds and algorithms for settings both with and without cluster-size augmentation. In particular, we prove that any deterministic online algorithm has a competitive ratio of at least $k$, even with significant augmentation. Our main algorithmic contributions are an $O(k \log{k})$-competitive deterministic algorithm for the general setting with constant augmentation, and a constant competitive algorithm for the maximum matching variant.

cs.DS↗

arXiv2020

Push-Down Trees: Optimal Self-Adjusting Complete Trees

This paper studies a fundamental algorithmic problem related to the design of demand-aware networks: networks whose topologies adjust toward the traffic patterns they serve, in an online manner. The goal is to strike a tradeoff between the benefits of such adjustments (shorter routes) and their costs (reconfigurations). In particular, we consider the problem of designing a self-adjusting tree network which serves single-source, multi-destination communication. The problem has interesting connections to self-adjusting datastructures. We present two constant-competitive online algorithms for this problem, one randomized and one deterministic. Our approach is based on a natural notion of Most Recently Used (MRU) tree, maintaining a working set. We prove that the working set is a cost lower bound for any online algorithm, and then present a randomized algorithm RANDOM-PUSH which approximates such an MRU tree at low cost, by pushing less recently used communication partners down the tree, along a random walk. Our deterministic algorithm MOVE-HALF does not directly maintain an MRU tree, but its cost is still proportional to the cost of an MRU tree, and also matches the working set lower bound.

cs.DS↗

arXiv2019

Measuring the Complexity of Packet Traces

This paper studies the structure of several real-world traces (including Facebook, High-Performance Computing, Machine Learning, and simulation generated traces) and presents a systematic approach to quantify and compare the structure of packet traces based on the entropy contained in the trace file. Insights into the structure of packet traces can lead to improved network algorithms that are optimized toward specific traffic patterns. We then present a methodology to quantify the temporal and non-temporal components of entropy contained in a packet trace, called the trace complexity, using randomization and compression. We show that trace complexity provides unique insights into the characteristics of various applications and argue that there is a need for traffic generation models that preserve the intrinsic structure of empirically measured application traces. We then propose a traffic generator model that is able to produce a synthetic trace that matches the complexity level of its corresponding real-world trace.

cs.NI↗

arXiv2019

Self-Adjusting Linear Networks

Emerging networked systems become increasingly flexible and reconfigurable. This introduces an opportunity to adjust networked systems in a demand-aware manner, leveraging spatial and temporal locality in the workload for online optimizations. However, it also introduces a trade-off: while more frequent adjustments can improve performance, they also entail higher reconfiguration costs. This paper initiates the formal study of linear networks which self-adjust to the demand in an online manner, striking a balance between the benefits and costs of reconfigurations. We show that the underlying algorithmic problem can be seen as a distributed generalization of the classic dynamic list update problem known from self-adjusting datastructures: in a network, requests can occur between node pairs. This distributed version turns out to be significantly harder than the classical problem in generalizes. Our main results are a $Ω(\log{n})$ lower bound on the competitive ratio, and a (distributed) online algorithm that is $O(\log{n})$-competitive if the communication requests are issued according to a linear order.

cs.DS↗

arXiv2019

ReNets: Toward Statically Optimal Self-Adjusting Networks

This paper studies the design of self-adjusting networks whose topology dynamically adapts to the workload, in an online and demand-aware manner. This problem is motivated by emerging optical technologies which allow to reconfigure the datacenter topology at runtime. Our main contribution is ReNet, a self-adjusting network which maintains a balance between the benefits and costs of reconfigurations. In particular, we show that ReNets are statically optimal for arbitrary sparse communication demands, i.e., perform at least as good as any fixed demand-aware network designed with a perfect knowledge of the future demand. Furthermore, ReNets provide compact and local routing, by leveraging ideas from self-adjusting datastructures.

cs.NI↗

arXiv2018

Toward Demand-Aware Networking: A Theory for Self-Adjusting Networks

The physical topology is emerging as the next frontier in an ongoing effort to render communication networks more flexible. While first empirical results indicate that these flexibilities can be exploited to reconfigure and optimize the network toward the workload it serves and, e.g., providing the same bandwidth at lower infrastructure cost, only little is known today about the fundamental algorithmic problems underlying the design of reconfigurable networks. This paper initiates the study of the theory of demand-aware, self-adjusting networks. Our main position is that self-adjusting networks should be seen through the lense of self-adjusting datastructures. Accordingly, we present a taxonomy classifying the different algorithmic models of demand-oblivious, fixed demand-aware, and reconfigurable demand-aware networks, introduce a formal model, and identify objectives and evaluation metrics. We also demonstrate, by examples, the inherent advantage of demand-aware networks over state-of-the-art demand-oblivious, fixed networks (such as expanders).

cs.NI↗

arXiv2018

Fault-Tolerant Hotelling Games

The $n$-player Hotelling game calls for each player to choose a point on the line segment, so as to maximize the size of his Voronoi cell. This paper studies fault-tolerant versions of the Hotelling game. Two fault models are studied: line faults and player faults. The first model assumes that the environment is prone to failure: with some probability, a disconnection occurs at a random point on the line, splitting it into two separate segments and modifying each player's Voronoi cell accordingly. A complete characterization of the Nash equilibria of this variant is provided for every $n$. Additionally, a one to one correspondence is shown between equilibria of this variant and of the Hotelling game with no faults. The second fault model assumes the players are prone to failure: each player is removed from the game with i.i.d. probability, changing the payoffs of the remaining players accordingly. It is shown that for $n \geq 3$ this variant of the game has no Nash equilibria.

cs.GT↗

arXiv2017

Towards Communication-Aware Robust Topologies

We currently witness the emergence of interesting new network topologies optimized towards the traffic matrices they serve, such as demand-aware datacenter interconnects (e.g., ProjecToR) and demand-aware overlay networks (e.g., SplayNets). This paper introduces a formal framework and approach to reason about and design such topologies. We leverage a connection between the communication frequency of two nodes and the path length between them in the network, which depends on the entropy of the communication matrix. Our main contribution is a novel robust, yet sparse, family of network topologies which guarantee an expected path length that is proportional to the entropy of the communication patterns.

cs.NI↗

arXiv2017

Demand-Aware Network Designs of Bounded Degree

Traditionally, networks such as datacenter interconnects are designed to optimize worst-case performance under arbitrary traffic patterns. Such network designs can however be far from optimal when considering the actual workloads and traffic patterns which they serve. This insight led to the development of demand-aware datacenter interconnects which can be reconfigured depending on the workload. Motivated by these trends, this paper initiates the algorithmic study of demand-aware networks (DANs) designs, and in particular the design of bounded-degree networks. The inputs to the network design problem are a discrete communication request distribution, D, defined over communicating pairs from the node set V , and a bound, d, on the maximum degree. In turn, our objective is to design an (undirected) demand-aware network N = (V,E) of bounded-degree d, which provides short routing paths between frequently communicating nodes distributed across N. In particular, the designed network should minimize the expected path length on N (with respect to D), which is a basic measure of the efficiency of the network. We show that this fundamental network design problem exhibits interesting connections to several classic combinatorial problems and to information theory. We derive a general lower bound based on the entropy of the communication pattern D, and present asymptotically optimal network-aware design algorithms for important distribution families, such as sparse distributions and distributions of locally bounded doubling dimensions.

cs.DC↗

arXiv2017

Assortative Mixing Equilibria in Social Network Games

It is known that individuals in social networks tend to exhibit homophily (a.k.a. assortative mixing) in their social ties, which implies that they prefer bonding with others of their own kind. But what are the reasons for this phenomenon? Is it that such relations are more convenient and easier to maintain? Or are there also some more tangible benefits to be gained from this collective behaviour? The current work takes a game-theoretic perspective on this phenomenon, and studies the conditions under which different assortative mixing strategies lead to equilibrium in an evolving social network. We focus on a biased preferential attachment model where the strategy of each group (e.g., political or social minority) determines the level of bias of its members toward other group members and non-members. Our first result is that if the utility function that the group attempts to maximize is the degree centrality of the group, interpreted as the sum of degrees of the group members in the network, then the only strategy achieving Nash equilibrium is a perfect homophily, which implies that cooperation with other groups is harmful to this utility function. A second, and perhaps more surprising, result is that if a reward for inter-group cooperation is added to the utility function (e.g., externally enforced by an authority as a regulation), then there are only two possible equilibria, namely, perfect homophily or perfect heterophily, and it is possible to characterize their feasibility spaces. Interestingly, these results hold regardless of the minority-majority ratio in the population. We believe that these results, as well as the game-theoretic perspective presented herein, may contribute to a better understanding of the forces that shape the groups and communities of our society.

cs.SI↗

arXiv2016

Information Spreading in Dynamic Networks under Oblivious Adversaries

We study the problem of gossip in dynamic networks controlled by an adversary that can modify the network arbitrarily from one round to another, provided that the network is always connected. In the gossip problem, $n$ tokens are arbitrarily distributed among the $n$ network nodes, and the goal is to disseminate all the $n$ tokens to every node. Our focus is on token-forwarding algorithms, which do not manipulate tokens in any way other than storing, copying, and forwarding them. Gossip can be completed in linear time in any static network, but a basic open question for dynamic networks is the existence of a distributed protocol that can do significantly better than an easily achievable bound of $O(n^2)$ rounds. In previous work, it has been shown that under adaptive adversaries, every token forwarding algorithm requires $Ω(n^2/\log n)$ rounds. In this paper, we study oblivious adversaries, which differ from adaptive adversaries in one crucial aspect--- they are oblivious to random choices made by the protocol. We present an $\tildeΩ(n^{3/2})$ lower bound under an oblivious adversary for RANDDIFF, a natural algorithm in which neighbors exchange a token chosen uniformly at random from the difference of their token sets. We also present an $\tildeΩ(n^{4/3})$ bound under a stronger notion of oblivious adversary for symmetric knowledge-based algorithms. On the positive side, we present a centralized algorithm that completes gossip in $\tilde{O}(n^{3/2})$ rounds. We also show an $\tilde{O}(n^{5/3})$ upper bound for RANDDIFF in a restricted class of oblivious adversaries, which we call paths-respecting.

cs.DC↗

arXiv2015

Breaking the log n barrier on rumor spreading

$O(\log n)$ rounds has been a well known upper bound for rumor spreading using push&pull in the random phone call model (i.e., uniform gossip in the complete graph). A matching lower bound of $Ω(\log n)$ is also known for this special case. Under the assumption of this model and with a natural addition that nodes can call a partner once they learn its address (e.g., its IP address) we present a new distributed, address-oblivious and robust algorithm that uses push&pull with pointer jumping to spread a rumor to all nodes in only $O(\sqrt{\log n})$ rounds, w.h.p. This algorithm can also cope with $F= O(n/2^{\sqrt{\log n}})$ node failures, in which case all but $O(F)$ nodes become informed within $O(\sqrt{\log n})$ rounds, w.h.p.

cs.DS↗

arXiv2015

Distributed Computing on Core-Periphery Networks: Axiom-based Design

Inspired by social networks and complex systems, we propose a core-periphery network architecture that supports fast computation for many distributed algorithms and is robust and efficient in number of links. Rather than providing a concrete network model, we take an axiom-based design approach. We provide three intuitive (and independent) algorithmic axioms and prove that any network that satisfies all axioms enjoys an efficient algorithm for a range of tasks (e.g., MST, sparse matrix multiplication, etc.). We also show the minimality of our axiom set: for networks that satisfy any subset of the axioms, the same efficiency cannot be guaranteed for any deterministic algorithm.

cs.DC↗

arXiv2015

Random Preferential Attachment Hypergraphs

The random graph model has recently been extended to a random preferential attachment graph model, in order to enable the study of general asymptotic properties in network types that are better represented by the preferential attachment evolution model than by the ordinary (uniform) evolution lodel. Analogously, this paper extends the random {\em hypergraph} model to a random {\em preferential attachment hypergraph} model. We then analyze the degree distribution of random preferential attachment hypergraphs and show that they possess heavy tail degree distribution properties similar to those of random preferential attachment graphs. However, our results show that the exponent of the degree distribution is sensitive to whether one considers the structure as a hypergraph or as a graph.

cs.SI↗

arXiv2014

Core-Periphery in Networks: An Axiomatic Approach

Recent evidence shows that in many societies worldwide the relative sizes of the economic and social elites are continuously shrinking. Is this a natural social phenomenon? What are the forces that shape this process? We try to address these questions by studying a Core-Periphery social structure composed of a social elite, namely, a relatively small but well-connected and highly influential group of powerful individuals, and the rest of society, the periphery. Herein, we present a novel axiom-based model for the forces governing the mutual influences between the elite and the periphery. Assuming a simple set of axioms, capturing the elite's dominance, robustness, compactness and density, we are able to draw strong conclusions about the elite-periphery structure. In particular, we show that a balance of powers between elite and periphery and an elite size that is sub-linear in the network size are universal properties of elites in social networks that satisfy our axioms. We note that the latter is in controversy to the common belief that the elite size converges to a linear fraction of society (most recently claimed to be 1%). We accompany these findings with a large scale empirical study on about 100 real-world networks, which supports our results.

cs.SI↗

arXiv2013

Simple Destination-Swap Strategies for Adaptive Intra- and Inter-Tenant VM Migration

This paper investigates the opportunities and limitations of adaptive virtual machine (VM) migration to reduce communication costs in a virtualized environment. We introduce a new formal model for the problem of online VM migration in two scenarios: (1) VMs can be migrated arbitrarily in the substrate network; e.g., a private cloud provider may have an incentive to reduce the overall communication cost in the network. (2) VMs can only be migrated within a given tenant; e.g., a user that was assigned a set of physical machines may exchange the functionality of the VMs on these machines. We propose a simple class of Destination-Swap algorithms which are based on an aggressive collocation strategy (inspired by splay datastructures) and which maintain a minimal and local amount of per-node (amortized cost) information to decide where to migrate a VM and how; thus, the algorithms react quickly to changes in the load. The algorithms come in two main flavors, an indirect and distributed variant which keeps existing VM placements local, and a direct variant which keeps the number of affected VMs small. We show that naturally, inter-tenant optimizations yield a larger potential for optimization, but generally also a tenant itself can improve its embedding. Moreover, there exists an interesting tradeoff between direct and indirect strategies: indirect variants are preferable under skewed and sparse communication patterns due to their locality properties.

cs.DC↗

arXiv2013

SINR Diagram with Interference Cancellation

In this paper we study the reception zones of a wireless network in the SINR model with receivers that employ \emph{interference cancellation} (IC), a technique that allows a receiver to decode interfering signals, and \emph{cancel} them from the received signal in order to decode its intended message. We first derive some important topological properties of the diagram describing the reception zones and their connections to \emph{high-order Voronoi diagrams} and other related geometric objects. We then discuss the computational issues that arise when seeking an efficient description of the zones. Our main fundamental result states that although potentially there are exponentially many possible cancellation orderings (and consequently reception cells), in fact there are much fewer nonempty such cells. We prove a (tight) linear bound on the number of cells and provide a polynomial time algorithm to describe the diagram. Moreover, we introduce a novel measure, referred to as the \emph{Compactness Parameter}, which influences the tightness of our bounds. We then utilize the properties established for reception diagrams to devise a logarithmic time algorithm for answering \emph{point-location} queries for networks with IC.

cs.CG↗

arXiv2013

OBST: A Self-Adjusting Peer-to-Peer Overlay Based on Multiple BSTs

The design of scalable and robust overlay topologies has been a main research subject since the very origins of peer-to-peer (p2p) computing. Today, the corresponding optimization tradeoffs are fairly well-understood, at least in the static case and from a worst-case perspective. This paper revisits the peer-to-peer topology design problem from a self-organization perspective. We initiate the study of topologies which are optimized to serve the communication demand, or even self-adjusting as demand changes. The appeal of this new paradigm lies in the opportunity to be able to go beyond the lower bounds and limitations imposed by a static, communication-oblivious, topology. For example, the goal of having short routing paths (in terms of hop count) does no longer conflict with the requirement of having low peer degrees. We propose a simple overlay topology Obst(k) which is composed of k (rooted and directed) Binary Search Trees (BSTs), where k is a parameter. We first prove some fundamental bounds on what can and cannot be achieved optimizing a topology towards a static communication pattern (a static Obst(k)). In particular, we show that the number of BSTs that constitute the overlay can have a large impact on the routing costs, and that a single additional BST may reduce the amortized communication costs from Omega(log(n)) to O(1), where n is the number of peers. Subsequently, we discuss a natural self-adjusting extension of Obst(k), in which frequently communicating partners are "splayed together".

cs.NI↗