Ikuo Nakagawa

Inferring POP-Level ISP Topology through End-to-End Delay Measurement

In this paper, we propose a new topology inference technique that aims to reveals how ISPs deploy their layer two and three networks at the POP level, without relying on ISP core network information such as router hops and domain names. This is because, even though most of previous works in this field leverage core network information to infer ISP topologies, some of our measured ISPs filter ICMP packets and do not allow us to access core network information through traceroute. And, several researchers point out that such information is not always reliable. So, to infer ISP core network topology without relying on ISP releasing information, we deploy systems to measure end-to-end communication delay between residential users, and map the collected delay and corresponding POP-by-POP paths. In our inference process, we introduce assumptions about how ISPs tend to deploy their layer one and two networks. To validate our methodology, we measure end-to-end communication delay of four nationwide ISPs between thirteen different cities in Japan and infer their POP-level topologies.

Dripcast - Architecture and Implementation of Server-less Java Programming Framework for Billions of IoT Devices

We propose “Dripcast,” a new server-less Java programming framework for billions of IoT (Internet of Things) devices. The framework makes it easy to develop device applications working with a cloud, that is, scalable computing resources on the Internet. The framework consists of two key technologies; (1) transparent remote procedure call (2) mechanism to read, write and process Java objects with scale-out style distributed datastore. A great benefit of the framework is that there is no need to write a server-side program nor a database code. A very simple client-side program is enough to work with the framework, to read, write or process Java objects on a cloud. The mechanism is infinitely scalable since it works with scale-out technologies. In this paper, we describe the concept and the architecture of the Dripcast framework. We also implement the framework and evaluate from two points of views, 1) from the view point of scalability about cloud resources, 2) from the view point of method call encapsulation overhead in client IoT devices.

DHT Extension of m-Cloud -- Scalable and Distributed Privacy Preserving Statistical Computation on Public Cloud

In the age of IoT (Internet of Things), tens of billions sensors or devices will be connected to cloud services. Although cloud environment is suitable for incredibly growing IoT services because of elasticity, scalability and efficiency for collecting and analyzing such data, protecting privacy on public cloud is a big issue. We proposed m-cloud, a distributed computation mechanism on multiple and independent public cloud environment. The mechanism enables to calculate various statistical indices from a large number of sensor data while it reduces the risk of revealing privacy information. We also propose an extension of the m-cloud computation mechanism so that it provides high scalability for infinitely increasing number of IoT devices by applying DHT (Distributed Hash Table) technique.

An efficient algorithm for finding the minimum norm point in the convex hull of a finite point set in the plane

The minimum norm point problem is to find the minimum norm point in the convex hull of given n points in the d-dimensional Euclidean space. In this note, we propose an efficient algorithm for the two dimensional case and evaluate its time complexity.

Design and evaluation of a P2MP MPLS-based hierarchical service management system

This paper proposes a point-to-multipoint (P2MP) Multi-Protocol Label Switching (MPLS) based hierarchical service management system. Traditionally, general management systems deployed in some service providers control MPLS label switched paths (LSPs) (e.g. RSVP-TE, LDP) and services (e.g. L2VPN, L3VPN and IP) separately. In order for the dedicated management systems for MPLS LSPs and services to cooperate with each other automatically, a hierarchical service management system has been proposed with the main focus on point-to-point (P2P) TE LSPs in MPLS path management. In the case where P2MP TE LSPs and services are deployed in MPLS networks, the dedicated management systems for P2MP TE LSPs and services must work together automatically. Therefore, this paper proposes a new automatic correlation between P2MP TE LSPs and each service. In particular, this paper proposes a new algorithm for a correlation between P2MP TE LSPs and multicast VPN services. The capacity and performance searching on both a sub-LSP and correlation tables are evaluated, and the results show this system is very scalable within real MPLS production networks. This system, with automatic correlation, appears to be deployable in real MPLS production networks.

A Hierarchical Service Management System for MPLS Network Services

This paper proposes a hierarchical service management system for MPLS network services. Traditionally, general management systems which have been deployed in some service providers control MPLS LSPs (e.g. RSVP-TE, LDP) and services (e.g. L2VPN, L3VPN and IP) separately. If a fault occurs in an MPLS network, the dedicated management system for MPLS LSPs can detect the fault and recognize the state of MPLS LSPs. However, it cannot detect the extent of the impact due to the fault in each service. Furthermore, its own inability to identify the affected customer means it takes some time to identify the affected customers, cooperating manually with the dedicated management system for services. Therefore, this paper proposes a new automatic correlation between MPLS LSPs and each service. In particular, this paper proposes a new algorithm for a correlation between RSVP-TE LSPs and L3VPN services. Simulations are conducted to evaluate the capacity on a correlation table and the performance searching on a correlation table, and results show this system is very scalable within real MPLS production networks. This system, with the automatic correlation, could be sufficiently deployed in real MPLS production networks.

Design and implementation of overlaying multi-homing architecture

As a result of the vast Internet growth, a multi-homing technology is being extensively utilized to keep connectivity to important networks. However, the traditional multi-homing architecture severely impacts routing performance of the Internet making it non-scalable. Furthermore, although it can enhance reachability in case of connectivity anomalies, it can not be used to enhance utilization because it can not assign in-coming traffic to multiple links. We propose a new multi-homing architecture which is based on the overlay networking model, providing solutions to these problems. Our architecture improves scalability for increasing multi-homing users and realizes a dynamic and flexible line selection for in-coming traffic, while retaining the original advantage of the multi-homing scheme, which is continuous connectivity to the Internet. Of course, our proposed architecture can be applied not only to IPv4 networks but also to IPv6 networks.

A Design and Implementation of Global Distributed POSIX File System on the Top of Multiple Independent Cloud Services

We propose an extension of distcloud-fs for achieving a global distributed POSIX file system in intercloud environment. The file system works in intercloud environment, that is, on the top of multiple independent cloud services. The important capability of the intercloud extension is that there is no need for users to take care of infrastructure layer, while the file system stores data into multiple cloud services, transparently and automatically. As the result, cloud users may have diversity about vendors and locations for data platform, as discussed in various intercloud research projects. In this paper, we describe the architecture of the distcloud-fs and its intercloud extension. We also describe design and implementation of the prototype and discuss some typical use cases of the file system with practical demonstrations.