Kwonyong Lee

A QoS Assured Network Service Chaining Algorithm in Network Function Virtualization Architecture

In the Network Function Virtualization (NFV) architecture, Network Service Chaining (NSC) is consisted in a certain order of network elements so that it can provide flexible network services to users. Due to the complexity of network infrastructure, creating a service chain requires high operation cost especially in carrier-grade network service providers and supporting stringent QoS requirements is also a challenge. Although several vendors provide various solutions for the NSC, there is only few information and the detailed algorithm or implementation logic is hidden. This paper presents an NSC algorithm in NFV that assures QoS from the perspective of service providers. In order to formulate NSC path selection problem, we apply the NP complete genetic algorithm. The evaluation results show that the proposed algorithm minimizes the operation cost of service providers by approximately 10.6% while the requested QoS targets is not violated.

A flow-based prediction scheme to manage resources in enterprise data centers

Virtualization is widely used to enhance the utilization of the resources in the enterprise data centers. Migration can solve the resource shortage problem of a physical machine in virtualized environment. If the future resource usage can be predicted, the resource can be managed more efficiently. Auto Regression prediction scheme is well-known for its simplicity and accuracy, but it requires complex coefficient calculations. It is known that workloads of enterprise data center tend to have a specific pattern. In this environment, simpler prediction technique can be used without loss of accuracy. This paper introduces a flow-based prediction scheme which forecasts future resource usage when the usage has a pattern. Through the performance evaluation, we show that the proposed scheme outperforms the AR prediction scheme in terms of accuracy and overheads.

A Dynamic Message-Aware Communication Scheduler for Ceph Storage System

With the proliferation of cloud computing technologies, the Ceph, a distributed object-based storage system has been an attractive alternative to building a storage backend due to its excellent performance, reliability, and scalability. As the storage system processes huge amount data and the network traffic generated from the cloud increases rapidly, designing a high-performance messenger in the storage system has created a lot of challenging issues. Although the async messenger, one of the Cephs messengers, is known to be efficient and flexible, it contains several performance problems due to its simple round-robin based scheduling scheme that assigns a connection to a worker thread without any consideration for the amount of workloads transferred through the connections. This causes the imbalance of worker threads and adversely affects the performance of the Ceph storage system. This paper proposes a dynamic message-aware communication scheduler for Ceph storage system that balances the workloads of worker threads based on the types of incoming messages, while avoiding unnecessary connection movements among worker threads. We use genetic algorithm (GA) to solve this problem and implement the proposed scheduling algorithm using the async messenger. The benchmarking results show that the proposed approach outperforms the original async messenger by as much as 12.5% under the same workload from clients and 24% under the random workloads from clients.

Performance Optimization of Communication Subsystem in Scale-Out Distributed Storage

Scale-out distributed storage systems have recently gained high attentions with the emergence of big data and cloud computing technologies. However, these storage systems sometimes suffer from performance degradation, especially when the communication subsystem is not fully optimized. The problem becomes worse as the network bandwidth and its corresponding traffic increase. In this paper, we first conduct an extensive analysis of communication subsystem in Ceph, an object-based scale-out distributed storage system. Ceph uses asynchronous messenger framework for inter-component communication in the storage cluster. Then, we propose three major optimizations to improve the performance of Ceph messenger. These include i) deploying load balancing algorithm among worker threads based on the amount of workloads, ii) assigning multiple worker threads (we call dual worker) per single connection to maximize the overlapping activity among threads, and iii) using multiple connections between storage servers to maximize bandwidth usage, and thus reduce replication overhead. The experimental results show that the optimized Ceph messenger outperforms the original messenger implementation up to 40% in random writes with 4K messages. Moreover, Ceph with optimized communication subsystem shows up to 13% performance improvement as compared to original Ceph.

A service path selection and adaptation algorithm for QoS assurance and load balancing in context-aware service overlay networks

As services are becoming more complicated, service providers face with tremendous difficulties in developing their services. In this environment, a lot of component functions are provided by multiple service providers, and the providers also desire to reduce costs for providing their services. A context-aware service overlay network CSON is a promising architecture to meet those demands by dynamically composing service paths and reconfiguring them based on various contexts. This paper proposes a service path selection and adaptation algorithm that assures quality of services QoS and balances the loads at the same time. The algorithm also dynamically adapts to an optimal service path when it detects a large amount of unexpected service loads over the service path. Since our algorithm is to solve a variation of multi-constrained path selection problem, which is known to be NP-complete, we formulate the problem using ant colony optimisation algorithm.

A CPU Overhead-Aware VM Placement Algorithm for Network Bandwidth Guarantee in Virtualized Data Centers

As server consolidations based on the virtualization techniques become popular and cloud services continue to grow rapidly, more and more data centers are being built to accommodate a number of virtual clusters running various workloads. Since these virtual clusters often share the resources provided by physical machines (PMs), it is more likely that the interferences between virtual machines (VMs) affect the performance of applications running on top of the virtual clusters. While a lot of studies have proposed different virtual machine placement algorithms to investigate this issue, the problem caused by network performance variability still remains as a challenging issue. Since they usually ignore the CPU overhead to process the communications between VMs, the network bandwidth allocated to a VM cannot be fully utilized when a PM has not enough CPU resources to cover the CPU overhead for VM networking functions. This results in unpredictable application performance running on the virtual clusters. This paper proposes a virtual machine placement algorithm that considers the CPU overhead incurred to reserve network bandwidth in a virtualized data center environment. In order to decide the CPU overhead necessary to guarantee the network bandwidth allocated to a VM, a performance model based on standard linear regression using the data collected from a real environment is used. By comparing the amount of CPU resource available in the driver domain with the CPU overhead obtained from the performance model, the proposed algorithm decides whether the network bandwidth for the VM can be provided or not and selects an appropriate location for VM placement. The benchmarking results show that the proposed algorithm guarantees the network bandwidth allocated to each VM without violations when the CPU resources are shared by multiple VMs.

An adaptive data transfer algorithm using block device reconfiguration in virtual MapReduce clusters

With the proliferation of cloud computing and virtual machine technologies, MapReduce applications are increasingly deployed in clouds to leverage the full potential of cloud computing environments. However, the MapReduce, which is generally used for processing large amount of data, suffers from the I/O virtualization overheads and resource competitions among virtual machines when it is run on virtual clouds. This paper proposes an adaptive data transfer algorithm in virtual MapReduce clusters. The proposed algorithm utilizes a block device reconfiguration scheme, where a block device attached to a virtual machine can be dynamically detached and reattached to other virtual machines hosted in the same physical machine. By reconfiguring the block devices, we can easily move files across different virtual machines located at the same physical machine without any network transfers between virtual machines. When the output of each map task is transferred to the reducer, this algorithm adaptively determines an appropriate transfer method between network transfer and block device reconfiguration based on current CPU utilization values and the data size for the transfer. Even in the case of data transfer between virtual machines across multiple physical machines, we can remove the transfer overheads between the virtual machine and the driver domain, which results in reducing the data transfer time and performance effects to other virtual machines in the shuffle phase. We have implemented our algorithm in Hadoop MapReduce. The benchmarking results show that the overheads incurred by transferring data from mapper virtual machines to reducer virtual machines are minimized and the execution times of MapReduce applications are shortened.

Remote Binder: Remote Procedure Call between Android Devices

As Internet of Things(IoT) has become one of the most rapidly growing market in the world, the number of embedded Android devices has increased. Therefore, it is necessary to set up an environment that connects and cooperates between the devices via network. The environment requires an ability not only to obtain information about other devices through a network but to control remote devices by invoking remote procedures. This paper suggests the Remote Binder, which is a method for remote procedure call between devices operating on Android platform. It invokes procedures of other Android devices without any revisions via network by extending the binder structure which is used for inter-process communication in Android.