Search

Recently, edge computing is getting attention as a new computing paradigm that is expected to achieve short-delay and high-throughput task offloading for large scale Internet-of-Things (IoT) applications. In edge computing, workload distribution is one of the most critical issues that largely influences the delay and throughput performance of edge clouds, especially in distributed Function-as-a-Service (FaaS) over networked edge nodes. In this paper, we propose the Resource Allocation Control Engine with Reinforcement learning (RACER), which provides an efficient workload distribution strategy to reduce the task response slowdown with per-task response time Quality-of-Service (QoS). First, we present a novel problem formulation with the per-task QoS constraint derived from the well-known token bucket mechanism . Second, we employ a problem relaxation to reduce the overall computation complexity by compromising just a bit of optimality. Lastly, we take the deep reinforcement learning approach as an alternative solution to the workload distribution problem to cope with the uncertainty and dynamicity of underlying environments. Evaluation results show that RACER achieves a significant improvement in terms of per-task QoS violation ratio, average slowdown, and control efficiency, compared to AREA, a state-of-the-art workload distribution method.

User mobility characteristics are considered to significantly affect the performance and UX of modern IT systems such as mobile communications, context-aware services, location-based applications, smart mobility, and so on. Due to privacy concerns, legal issues, and expensive measurement costs for real user mobility traces, synthetic mobility traces are widely used for research and development purposes. Until now, most of the user mobility synthesis approaches have been categorized into two major paradigms: stochastic modeling and simulation. However, they naturally have a limitation in mimicking actual user movements because it is impossible to capture all temporal, spatial, and behavioral characteristics that are implicitly melted in real user mobility. Along with the recent advancements of deep learning technologies, there have been several challenging proposals that exploit the power of the deep learning, especially generative models, in mobility synthesis. This paper reviews and summarizes recently proposed user mobility synthesis schemes based on generative adversarial networks that have been one of the most leading deep learning technologies for the last few years.

The performance of Functions-as-a-Service (FaaS) would be significantly improved by organizing cloud servers into a hierarchical distributed architecture, resulting in low-latency access and faster response when compared to centralized cloud. However, the distributed organization introduces a new type of decision making problem for placing and executing functions to a specific cloud server. In order to handle the problem, we extended a well-known cloud computing simulator, CloudSim. The extended CloudSim enables users to define FaaS functions with various characteristics and service level objectives (SLOs), place them across geo-distributed cloud servers, and evaluate per-function performance. Proof-of-Concept (PoC) evaluation results show the potential of our CloudSim extension in terms of execution efficiency and simulation reality.

With the advent of cloud and virtualization technologies and the integration of various computer communication technologies, today's computing environments can provide virtualized high quality services. The network traffic has also continuously increased with remarkable growth. Software defined networking/network function virtualization (SDN/NFV) enhancing the infrastructure agility, thus network operators and service providers are able to program their own network functions on vendor independent hardware substrate. However, in order for the SDN/NFV to realize a profit, it must provide a new resource sharing and monitoring procedures among the regionally distributed and virtualized computers. In this paper, we proposes a NFV monitoring architecture based practical measuring framework for network performance measurement. We also proposes a end-to-end connectivity support platform across a whole SDN/NFV networks has not been fully addressed.

Current information & communication technology infrastructures are struggling for a wide variety of devices, services, and business and technology evolution. Cloud computing is becoming a promising platform and has developed rapidly. In contrast to traditional enterprise IT solution, it enables enterprises to procure computing resources on demand basis and delegate management of all the resources to the cloud service provider. Healthcare information networks have emerged as one of the major research area in the cloud networks. Healthcare information is communication and information systems and technology that facilitate quality patient care, progressive medical education, and innovative research. Healthcare information networks must ensure the reliability and efficiency because it transfers highly personal data. Thus, it guarantees and keeps a certain level of transfer efficiency. In the distributed cloud environments, management costs for the network and computing resources are solved fundamentally through the integrated management system. It can increase the cost savings to solve the traffic explosion problem of core network via a distributed micro datacenters (DCs). However, traditional flooding methods may cause a lot of traffic because it had to send data to all the neighbor DCs. Restricted Path Flooding algorithms have been proposed for this purpose. In large networks, there is still the disadvantage that may occur traffic. In this paper, we develope lightweight path flooding algorithm to improve existing flooding algorithm using hop count restriction. This paper also investigates the problem of partial information sharing.

With the cloud computing, storage and computing resources are moving to remote resources such as virtual servers and storage systems in large datacenters, which raise many performance and management challenges. The distributed cloud networks must have an ability to maintain an acceptable level of service during network failures to provide high resilience against service failures. This paper presents an eccentric fault tolerant path grouping scheme that can achieve mobility independent Quality of Service guaranteed services in distributed cloud networks To make this possible, we propose a path allocation and restoration strategy called "Reverse Shared Risk Link Group" and demonstrates how this concept can be applied to minimize recovery contention in cloud networks and to assess the network convergence associated with failures. To reflect little more actual situations, we assume that the bandwidth of cloud connection requests can be some fraction of the lightpath capacity. Simulation results show that the proposed recovery scheme outperforms other schemes with fewer signaling traffics and contention probability. Very fast restorability can be accomplished for multiple failures. We also has developed an analytical model and performed analysis for the proposed scheme in terms of two performance factors: mean system time and reservation blocking probability.

With the rise of smart services, new services are no longer generated via the traditional linear models of innovation. Rather, smart-service innovations require new frameworks to integrate the market-pull and technology-push models, which focus on both user-centric considerations and technological developments. However, current streams of research continue to merely suggest a conceptual framework. Therefore, it appears to be difficult to apply this framework into practice, due to a lack of systematic methodologies. In response, this study employs morphological analysis, a powerful tool for the systematic generation of ideas. Since the morphological analysis employs several dimensions for explaining new services and combines them to generate new ideas, it is easy to integrate the different approaches to innovations, such as market pull and technology push. A case study of smartphone application services is also conducted.

The data transferring on health-care network has a sensitive data such as personal medical information, genome, and so on. So it is very important to make secure network, and there are potential targets to infect when the next vulnerability announces. And hacking or worm virus uses several vulnerability to drain important information. Also it attacks security program for personal computer such as vaccine, personal firewall, etc. Recently network fault is often happened by worm virus, and then the major reason is anomaly traffic flowed from inside host rather than attack from outside. In this paper, we try to design and implement automatic system which can trace the location for the infected system using monitoring load of network node and detecting signs of anomaly traffic.

As a large-scale content distribution becomes an important challenge in Internet, there are developed various approaches to address the challenge. ITU-T has standardized SUN (Smart Ubiquitous Network) as an evolved network concept that handles emerging Internet issues and now they are standardizing a functional architecture for content delivery in SUN. In this paper, we provide an overview of the functional architecture and operational procedures for content delivery in SUN.

Non-optimal path and IP-in-IP encapsulation in Mobile IP cause that normal RSVP does not work well with Mobile IP. Application mobility using SIP can be a counterproposal to Mobile IP in terms of above discordance with RSVP. However, its address allocation delay causes a considerable handoff latency compared to Mobile IP. PAR-SIP application mobility is useful to reduce the handoff latency by reserving a prospective address. Also PAR-SIP proactive handoff allows RSVP modules to reserve resources in advance. This paper presents a resource reservation scheme with PAR-SIP soft handoff to achieve Quality of Service (QoS) of real-time multimedia communications.

MRSVP and HMRSVP are representative RSVP extensions for mobile Internet. They reserve resources in advance to expedite the RSVP process and adopt RSVP Tunnel to avoid the RSVP message invisibility problem in Mobile IP. However, they still suffer from issues such as excessive advanced reservations and bottleneck problem due to RSVP Tunnel. In this paper, we evaluate the each performance of MRSVP and HMRSVP via simulation, and analyze the simulation results in terms of above issues.

IEEE 802.16 is one of broadband wireless access technologies and will become a key technology for growing wireless IP network. However, Point-to-Point oriented network such as IEEE 802.16 introduces some issues in applying conventional standard IPv6 Neighbor Discovery Protocol (NDP). Following the previous 3GPP model for IPv6 NDP is not flexible in IPv6 link configuration since it assigns different network prefixes to each subscriber stations. We presents a mechanism which can allocate a common network prefix to all subscriber stations under the same IPv6 link. Through this mechanism, the standard IPv6 NDP can be applied to IEEE 802.16 networks without modifying conventional host-side operation.

This paper addresses a network side solution of fast handover in mobile IPv6. Existing FMIPv6 basically assumes the involvement of mobile nodes in handover procedure. However, in some cases, it may not be easy to realize fast handover function into all mobile nodes. In the case, a network operator may want to provide fast handover to users by using only networks entities. To achieve network side fast handover, this paper basically uses bi-directional tunnel between a previous access router and new access router. The proposed fast handover scheme has simpler procedure than the existing FMIPv6 mostly uses the existing messages.