Joon-Myung Kang

Application-driven bandwidth guarantees in datacenters

Providing bandwidth guarantees to specific applications is becoming increasingly important as applications compete for shared cloud network resources. We present CloudMirror, a solution that provides bandwidth guarantees to cloud applications based on a new network abstraction and workload placement algorithm. An effective network abstraction should enable applications to easily and accurately specify their requirements, while simultaneously enabling the infrastructure to provision resources efficiently for deployed applications. Prior research has approached the bandwidth guarantee specification by using abstractions that resemble physical network topologies. We present a contrasting approach of deriving a network abstraction based on application communication structure, called Tenant Application Graph or TAG. CloudMirror also incorporates a new workload placement algorithm that efficiently meets bandwidth requirements specified by TAGs while factoring in high availability considerations. Extensive simulations using real application traces and datacenter topologies show that CloudMirror can handle 40% more bandwidth demand than the state of the art (e.g., the Oktopus system), while improving high availability from 20% to 70%.

PGA: Using Graphs to Express and Automatically Reconcile Network Policies

Software Defined Networking (SDN) and cloud automation enable a large number of diverse parties (network operators, application admins, tenants/end-users) and control programs (SDN Apps, network services) to generate network policies independently and dynamically. Yet existing policy abstractions and frameworks do not support natural expression and automatic composition of high-level policies from diverse sources. We tackle the open problem of automatic, correct and fast composition of multiple independently specified network policies. We first develop a high-level Policy Graph Abstraction (PGA) that allows network policies to be expressed simply and independently, and leverage the graph structure to detect and resolve policy conflicts efficiently. Besides supporting ACL policies, PGA also models and composes service chaining policies, i.e., the sequence of middleboxes to be traversed, by merging multiple service chain requirements into conflict-free composed chains. Our system validation using a large enterprise network policy dataset demonstrates practical composition times even for very large inputs, with only sub-millisecond runtime latencies.

Usage pattern analysis of smartphones

Recently, mobile traffic has increased tremendously due to the deployment of smart devices such as smartphones and smart tablets. These devices use various types of access networks such as 3G, WiFi, and mobile WiMAX. Network service providers also provide these access networks with various types of plans. There is a growing need to manage these smart devices and mobile networks. However, research on mobile network management has focused on the performance of the network itself. Few research has focused on applying the usage patterns of smartphone users to mobile network management. In this paper, we present an analysis of smartphone usage patterns. We define the five possible states of a smartphone based on such a phones basic operations. We collected real usage log data from real smartphone users over a two month period. We show that all users have their own usage pattern. We present a case study in order to show how to apply usage pattern information to power management of smartphones. We also discuss how to apply such information to mobile device management and network management.

Application-driven Bandwidth Guarantees in Datacenters

Providing bandwidth guarantees to specific applications is becoming increasingly important as applications compete for shared cloud network resources. We present CloudMirror, a solution that provides bandwidth guarantees to cloud applications based on a new network abstraction and workload placement algorithm. An effective network abstraction should enable applications to easily and accurately specify their requirements, while simultaneously enabling the infrastructure to provision resources efficiently for deployed applications. Prior research has approached the bandwidth guarantee specification by using abstractions that resemble physical network topologies. We present a contrasting approach of deriving a network abstraction based on application communication structure, called Tenant Application Graph or TAG. CloudMirror also incorporates a new workload placement algorithm that efficiently meets bandwidth requirements specified by TAGs while factoring in high availability considerations. Extensive simulations using real application traces and datacenter topologies show that CloudMirror can handle 40% more bandwidth demand than the state of the art (e.g., the Oktopus system), while improving high availability from 20% to 70%.

Personalized Battery Lifetime Prediction for Mobile Devices based on Usage Patterns

Nowadays mobile devices are used for various applications such as making voice/video calls, browsing the Internet, listening to music etc. The average battery consumption of each of these activities and the length of time a user spends on each one determines the battery lifetime of a mobile device. Previous methods have provided predictions of battery lifetime using a static battery consumption rate that does not consider user characteristics. This paper proposes an approach to predict a mobile device’s available battery lifetime based on usage patterns. Because every user has a different pattern of voice calls, data communication, and video call usage, we can use such usage patterns for personalized prediction of battery lifetime. Firstly, we define one or more states that affect battery consumption. Then, we record time-series log data related to battery consumption and the use time of each state. We calculate the average battery consumption rate for each state and determine the usage pattern based on the time-series data. Finally, we predict the available battery time based on the average battery consumption rate for each state and the usage pattern. We also present the experimental trials used to validate our approach in the real world. Category: Embedded computing

Autonomic personalized handover decisions for mobile services in heterogeneous wireless networks

In this paper, we present an autonomic management method to provide personalized handover decisions for customized mobility management in heterogeneous wireless networks. A handover decision is a significant problem, especially in a heterogeneous network environment. This is exacerbated when the goal is to provide personalized services for mobile users. Personalized handover decisions should not only consider received signal strength, which is a traditional handover decision factor, but also context information, user preferences, user profiles, and other non-functional requirements. We present two metrics for evaluating access points: access point acceptance value and access point satisfaction value. Our algorithm uses a combination of functional and non-functional metrics to select the access point that has the maximum satisfaction value. In our simulation study, we show that our decision algorithm is better than other decision algorithms in terms of end user satisfaction.

Software-Defined Infrastructure and the SAVI Testbed

In this paper we consider Software-Defined Infrastructure (SDI), a new concept for integrated control and management of converged heterogeneous resources. SDI enables programmability of infrastructure by enabling the support of cloud-based applications, customized network functions, and hybrid combinations of these. We motivate SDI in the context of a multi-tier cloud that includes massive-scale datacenters as well as a smart converged network edge. In SDI, a centralized SDI manager controls converged heterogeneous resources (i.e., computing, programmable hardware, and networking resources) using virtualization and a topology manager that provides the status of all resources and their connectivity. We discuss the design and implementation of SDI in the context of the Canadian SAVI testbed. We describe the current deployment of the SAVI testbed and applications that are currently supported in the testbed.

Towards autonomic handover decision management in 4g networks

Mobility management has become an important issue in 4G networks due to the integration of multiple network access technologies. Traditionally, only the received signal strength has been considered for the vertical handover. However, more considerations will be necessary to satisfy the end users preferences. In this paper, we propose an Autonomic Handover Manager (AHM) based on the autonomic computing concept to decide the best network interface to handover in 4G networks. AHM decides the appropriate policy for the specific service or application without the users intervention using the context information from the mobile terminal, the network and the user. We present the context information and the context evaluation function to decide handover based on the user preferences. We then describe the scenario to validate its feasibility using multimedia conferencing service on the mobile terminal.

Software-defined infrastructure and the Future Central Office

This paper discusses the role of virtualization and software-defined infrastructure (SDI) in the design of future application platforms, and in particular the Future Central Office (CO). A multi-tier computing cloud is presented in which resources in the Smart Edge of the network play a crucial role in the delivery of low-latency and data-intensive applications. Resources in the Smart Edge are virtualized and managed using cloud computing principles, but these resources are more diverse than in conventional data centers, including programmable hardware, GPUs, etc. We propose an architecture for future application platforms, and we describe the SAVI Testbed (TB) design for the Smart Edge. The design features a novel Software-Defined Infrastructure manager that operates on top of OpenStack and OpenFlow. We conclude with a discussion of the implications of the Smart Edge design on the Future CO.

User-Centric Prediction for Battery Lifetime of Mobile Devices

Today, mobile devices are being used for various applications such as making voice/video calls, browsing Internet and so on. The operating time and battery consumption spent in those activities affect the battery life of mobile devices. In this paper, we propose a method for predicting the battery lifetime of mobile devices based on usage patterns. We define the possible states of mobile devices based on their operating functions and develop a method of predicting battery lifetime based on average battery consumption and duration of each state.