We present \sys, a declarative, unsupervised root-cause analysis engine that uses time series monitoring data from large complex systems such as data centres. \sys empowers operators to succinctly specify a large number of causal hypotheses to search for causes of interesting events. \sys then ranks these hypotheses, reducing the number of causal dependencies from hundreds of thousands to a handful for human understanding. We show how a declarative language, such as SQL, can be effective in declaratively enumerating hypotheses that probe the structure of an unknown probabilistic graphical causal model of the underlying system. Our thesis is that databases are in a unique position to enable users to rapidly explore the possible causal mechanisms in data collected from diverse sources. We empirically demonstrate how \sys had helped us resolve over 30~performance issues in a commercial product since late 2014, of which we discuss a few cases in detail.

In this paper, we study the wireline two-unicast-Z communication network over directed acyclic graphs. The two-unicast-$Z$ network is a two-unicast network where the destination intending to decode the second message has a priori side information of the first message. We make three contributions in this paper. First, we describe a new linear network coding algorithm for two-unicast-Z networks over the directed acyclic graphs. Our approach includes the idea of interference alignment as one of its key ingredients. For the graphs of a bounded degree, our algorithm has linear complexity in terms of the number of vertices, and the polynomial complexity in terms of the number of edges. Second, we prove that our algorithm achieves the rate pair (1, 1) whenever it is feasible in the network. Our proof serves as an alternative, albeit restricted to two-unicast-Z networks over the directed acyclic graphs, to an earlier result of Wang et al., which studied the necessary and sufficient conditions for the feasibility of the rate pair (1, 1) in two-unicast networks. Third, we provide a new proof of the classical max-flow min-cut theorem for the directed acyclic graphs.

We derive a new linear network coding algorithm for two-unicast-Z networks over directed acyclic graphs, that is, for two-unicast networks where one destination has apriori information of the interfering source message. Our algorithm discovers linear network codes for two-unicast-Z networks by combining ideas of random linear network coding and interference neutralization. We show that our algorithm outputs an optimal network code for networks where there is only one edge emanating from each of the two sources. The complexity of our algorithm is polynomial in the number of edges of the graph.

We study two-unicast-Z networks1- two-source two-destination (two-unicast) wireline networks over directed acyclic graphs, where one of the two destinations (say the second destination) is apriori aware of the interfering (first) source's message. For certain classes of two-unicast-Z networks, we show that the rate-tuple (N, 1) is achievable as long as the individual source-destination cuts for the two source-destination pairs are respectively at least as large as N and 1, and the generalized network sharing cut - a bound previously defined by Kamath et. al. - is at least as large as N +1. We show this through a novel achievable scheme which is based on random linear coding at all the edges in the network, except at the GNS-cut set edges, where the linear coding co-efficients are chosen in a structured manner to cancel interference at the receiver first destination.

Existing mobile devices have the capability to use multiple network technologies simultaneously to help increase performance; but they rarely, if at all, effectively use these technologies in parallel. We first present empirical data to help understand the mobile environment when three heterogeneous networks are available to the mobile device (i.e., a WiFi network, WiMax network, and an Iridium satellite network). We then propose a reliable, multi-path protocol called Multi-Path TCP with Network Coding (MPTCP/NC) that utilizes each of these networks in parallel. An analytical model is developed and a mean-field approximation is derived that gives an estimate of the protocol's achievable throughput. Finally, a comparison between MPTCP and MPTCP/NC is presented using both the empirical data and mean-field approximation. Our results show that network coding can provide users in mobile environments a higher quality of service by enabling the use of multiple network technologies and the capability to overcome packet losses due to lossy, wireless network connections.

Throughput and per-packet delay can present strong trade-offs that are important in the cases of delay sensitive applications. We investigate such trade-offs using a random linear network coding scheme for one or more receivers in single hop wireless packet erasure broadcast channels. We capture the delay sensitivities across different types of network applications using a class of delay metrics based on the norms of packet arrival times. With these delay metrics, we establish a unified framework to characterize the rate and delay requirements of applications and to optimize system parameters. In the single receiver case, we demonstrate the trade-off between average packet delay, which we view as the inverse of throughput, and maximum inorder inter-arrival delay for various system parameters. For a single broadcast channel with multiple receivers having different delay constraints and feedback delays, we jointly optimize the coding parameters and time-division scheduling parameters at the transmitter. We formulate the optimization problem as a Generalized Geometric Program (GGP). This approach allows the transmitter to adjust adaptively the coding and scheduling parameters for efficient allocation of network resources under varying delay constraints. In the case where the receivers are served by multiple non-interfering wireless broadcast channels, the same optimization problem is formulated as a Signomial Program, which is NP-hard in general. We provide approximation methods using successive formulation of geometric programs and show the convergence of approximations.