Network functions and service function chaining are prevalent in cloud and ISP networks. In traditional software-based solutions, scaling up the capacity of these functions requires a large number of server cores. However, edge clouds are severely resource-constrained in terms of space, power, and budget, so traditional methods incur a high cost. We present Dejavu, a system that can offload a service chain to a programmable switch to achieve high performance and resource efficiency. Our system can compose multiple network functions into a single program that preserves the original chaining requirements, and exploit features of the switch ASIC to efficiently deploy the composed program on a single switch.

Data multicast is a crucial data transfer pattern in distributed big-data processing. However, due to the lack of network and system level support, data processing relies on unicast-based application layer multicast. In recent years, there has been a surge in interest in using various emerging circuit switching technologies to build data centers having hybrid packet-circuit switched rack-level interconnections, i.e., hybrid data centers. These physical layer innovations fundamentally change the inter-rack communication capability, especially the capability of multicast communication. We propose Republic, a complete system that addresses the challenging issues in achieving high-performance data multicast in hybrid data centers. Republic abstracts the underlying network complexity as a data multicast service and provides a unified Republic API for data center applications requesting data multicast. Republic is implemented and deployed in a hybrid data center testbed. Testbed evaluation shows that Republic can improve data multicast in Apache Spark machine learning applications by as much as 4.0x.

Shareable backup is an economical and effective way to mask failures from application performance. A small number of backup switches are shared network-wide for repairing failures on demand so that the network quickly recovers to its full capacity without applications noticing the failures. This approach avoids complications and ineffectiveness of rerouting. We propose ShareBackup as a prototype architecture to realize this concept and present the detailed design. We implement ShareBackup on a hardware testbed. Its failure recovery takes merely 0.73ms, causing no disruption to routing; and it accelerates Spark and Tez jobs by up to 4.1X under failures. Large-scale simulations with real data center traffic and failure model show that ShareBackup reduces the percentage of job flows prolonged by failures from 47.2% to as little as 0.78%. In all our experiments, the results for ShareBackup have little difference from the no-failure case.

This paper promotes convertible data center network architectures, which can dynamically change the network topology to combine the benefits of multiple architectures. We propose the flat-tree prototype architecture as the first step to realize this concept. Flat-tree can be implemented as a Clos network and later be converted to approximate random graphs of different sizes, thus achieving both Clos-like implementation simplicity and random-graph-like transmission performance. We present the detailed design for the network architecture and the control system. Simulations using real data center traffic traces show that flat-tree is able to optimize various workloads with different topology options. We implement an example flat-tree network on a 20-switch 24-server testbed. The traffic reaches the maximal throughput in 2.5s after a topology change, proving the feasibility of converting topology at run time. The network core bandwidth is increased by 27.6% just by converting the topology from Clos to approximate random graph. This improvement can be translated into acceleration of applications as we observe reduced communication time in Spark and Hadoop jobs.

We study the problem of supporting simultaneous query in WSNs. For such networks, the network efficiency can be significantly improved if the poller can obtain information via a simultaneous query. Two fundamental cases of such queries are studied in this paper: counting and identifying active neighboring nodes. We propose two mechanisms, Power based counting (Poc) and Power based identification (Poid), which achieve the goals by allowing neighbors to respond simultaneously to a poller. A key observation that motivates our design is that the power of a superposed signal increases with the number of component signals under the condition that the component signals are synchronized precisely. However, such high precision of synchronization is rather difficult to achieve in WSNs. To address this challenge, we design delicate delay compensation methods to reduce the phase offset of each component signal. Moreover, we propose a novel probabilistic estimation technique to overcome the hardware limitations on the observed received power. Poid currently works only in sparse networks though, we discuss and analyze the time complexity of applying Poid in dense networks. Experimental results show that the average accuracy of Poc and Poid is above 95 and 91 percent, respectively. In addition, our methods achieve substantially lower energy consumption and estimation delay compared with the state-of-the-art solutions.

Flooding is one of the most fundamental functions in wireless sensor networks (WSNs) and has been investigated extensively. Existing flooding schemes in duty-cycle WSNs can be categorized into two categories: synchronous ones and asynchronous ones. In practice asynchronous schemes are preferable since synchronous ones introduce more complexity and overhead for necessary clock synchronization. Existing asynchronous approaches are however imperfect for reliable flooding in duty-cycle WSNs. For example, opportunistic flooding is less flexible and Asynchronous Duty-cycle Broadcasting (ADB) suffers unsatisfactory flooding latency. We propose Constructive Interference-based Reliable Flooding (CIRF) in this paper, a novel design for reliable flooding in asynchronous duty-cycle WSNs. CIRF is integrated with the MAC protocol Receiver-Initiated MAC (RI-MAC) to improve the utilization of wireless medium and guarantee one-hop reliable transmission. The key idea of CIRF is to exploit the constructive interference feature when concurrent transmission occurs, which can be common in RI-MAC based WSNs. Simulation results indicate that CIRF achieves reliable flooding with reduced flooding latency, higher energy efficiency and delivery ratio compared to existing schemes.

Understanding the interests and behaviors of Web users serves as the core of many Web usage data mining applications such as personalized search, recommendation, personalization, business decision, advertisement targeting, marketing and customer relationship management. In this paper, an approach of building an adaptive user profile based on memory-model is proposed. The adaptive user profile gives a quantitative measurement of user interests and their changes, which reflects and predicts the interests dynamically and quantitatively, such as the enhancing, decaying of interests and the new interests coming out. It can differentiate long-term and short-term interests as well. It is applied to rank the topic list in which users are interested. In 17160 cases of user log, 70.94% of the nDCG results between user interest ranking list based on memory model and actual ranking list of user are above 90%, significantly higher than reference modeling's.