Makoto Imase

GridFTP-APT: automatic parallelism tuning mechanism for data transfer protocol GridFTP

GridFTP has been used as a data transfer protocol to effectively transfer a large volume of data in grid computing. GridFTP supports a feature called parallel data transfer that improves throughput by establishing multiple TCP connections in parallel. However, for achieving high GridFTP throughput, the number of TCP connections should be optimized based on the network status. In this paper, we propose an automatic parallelism tuning mechanism called GridFTP-APT (GridFTP with automatic parallelism tuning) that adjusts the number of parallel TCP connections only using information measurable in the grid middleware. Through simulation experiments, we demonstrate that GridFTP-APT significantly improves the performance of GridFTP in various network environments.

Scalable and Efficient Ant-Based Routing Algorithm for Ad-Hoc Networks

SUMMARY Ants-based routing algorithms have attracted the attention of researchers because they are more robust, reliable, and scalable than other conventional routing algorithms. Since they do not involve extra message exchanges to maintain paths when network topology changes, they are suitable for mobile ad-hoc networks where nodes move dynamically and topology changes frequently. As the number of nodes increases, however, the number of ants (i.e., mobile agents or control messages) also increases, which means that existing algorithms have poor scalability. In this paper, we propose a scalable ant-based routing algorithm that keeps the overhead low while keeping paths short. Our algorithm uses a multistep TTL (Time To Live) scheme, an effective message migration scheme, and an efficient scheme for updating the probability of packet forwarding. Simulation experiments have confirmed that our proposed algorithm can establish shorter paths than the conventional ant-based algorithm with the same signaling overhead.

On parameter tuning of data transfer protocol GridFTP for wide-area grid computing

In wide-area grid computing, geographically distributed computational resources are connected for enabling efficient and large-scale scientific/engineering computations. In the wide-area grid computing, a data transfer protocol called GridFTP has been commonly used for large file transfers. GridFTP has the following features for solving problems of the existing TCP. First, for accelerating the start-up in TCPs slow start phase and achieving high throughput in TCPs congestion avoidance phase, multiple TCP connections can be established in parallel. Second, according to the bandwidth-delay product of a network, the TCP socket buffer size can be negotiated between GridFTP server and client. However, in the literature, sufficient investigation has not been performed either on the optimal number of TCP connections or the optimal TCP socket buffer size. In this paper, we therefore quantitatively investigate the optimal parameter configuration of GridFTP in terms of the number of TCP connections and the TCP socket buffer size. We first derive performance metrics of GridFTP in steady state (i.e., goodput and packet loss probability). We then derive the optimal parameter configuration for GridFTP and quantitatively show performance limitations of GridFTP through several numerical examples. We also demonstrate validity of our approximate analysis by comparing simulation results with analytic ones

Understanding TCP over TCP : Effects of TCP tunneling on end-to-end throughput and latency

TCP tunnel is a technology that aggregates and transfers packets sent between end hosts as a single TCP connection. By using a TCP tunnel, the fairness among aggregated flows can be improved and several protocols can be transparently transmitted through a firewall. Currently, many applications such as SSH, VTun, and HTun use a TCP tunnel. However, since most applications running on end hosts generally use TCP, two TCP congestion controls (i.e., end-to-end TCP and tunnel TCP) operate simultaneously and interfere each other. Under certain conditions, it has been known that using a TCP tunnel severely degrades the end-to-end TCP performance. Namely, it has known that using a TCP tunnel drastically degrades the end-to-end TCP throughput for some time, which is called TCP meltdown problem. On the contrary, under other conditions, it has been known that using a TCP tunnel significantly improves the end-to-end TCP performance. However, it is still an open issue --- how, when, and why is a TCP tunnel malicious for end-to-end TCP performance? In this paper, we therefore investigate effect of TCP tunnel on end-to-end TCP performance using simulation experiments. Specifically, we quantitatively reveal effects of several factors (e.g., the propagation delay, usage of SACK option, TCP socket buffer size, and sender buffer size of TCP tunnel) on performance of end-to-end TCP and tunnel TCP.

On scalable modeling of TCP congestion control mechanism for large-scale IP networks

In this paper, we propose an analytic approach of modeling a closed-loop network with multiple feedback loops using fluid-flow approximation. Specifically, we model building blocks of a network (i.e., the congestion control mechanism of TCP, propagation delay of a transmission link, and the buffer of a router) as independent continuous-time systems. By interconnecting these systems, we obtain the model for a complex closed-loop network. We improve the accuracy of analytic models for TCP congestion control and RED router by extending existing fluid-flow models. First, we obtain a block diagram for each continuous-time system using a standard CAD tool widely used in control engineering. Second, we evaluate the performance of a closed-loop network with multiple feedback loops by connecting these block diagrams. We also validate the effectiveness of our analytic approach by comparing our analytic results with simulation results. Unlike other fluid-based modeling approaches, our analytic approach is scalable and accurate; our analytic approach is scalable in terms of the number of TCP connections and routers since both input/output of all continuous-time systems are uniformly defined as a packet transmission rate. Our analytic approach is accurate since the timeout mechanism of TCP and the packet dropping algorithm of RED router are rigorously modeled in our continuous-time systems.

On dynamic resource management mechanism using control theoretic approach for wide-area grid computing

In recent years, grid computing that integrates geographically distributed computing resources through communication networks captures the spotlight. Unlike parallel computing using conventional cluster computer systems, wide-area grid computing must resolve the following issues for effectively utilizing geographically distributed computing resources. First, since computing resources are shared by multiple users, the amount of available resources in a site changes over time. Second, since sites are geographically distributed, the transfer delay between sites cannot be neglected for computing resource management. For utilizing computing resources effectively, the amount of jobs injected into a site must be dynamically controlled according to the dynamically changing amount of available resources in sites. However, the transfer delay of a network is significant, so that it is not trivial for a resource allocation controller to quickly and dynamically adopt to the change in the amount of available resources in sites. In this paper, a dynamic resources management mechanism for wide-area grid computing called DRM-DC (dynamic resource management with delay compensator) based on control theory is proposed. The main feature of our DRM-DC is that it realizes high steady state and transient state performance in wide-area grid computing using a delay compensator called Smith predictor. Moreover, several discrete-time simulations using a Simgrid simulator are performed, and the effectiveness of our DRM-DC is demonstrated

Automatic parameter configuration mechanism for data transfer protocol GridFTP

In this paper, we propose an automatic parameter configuration mechanism for GridFTP, which optimizes the number of parallel TCP connections by utilizing analytic results in the work of Ito et al. (2005). The proposed mechanism first measures the network status (e.g., the goodput and the round-trip time of GridFTP data channels) at the GridFTP client. Based on these measurement results, it adjusts the number of parallel TCP connections for maximizing the GridFTP goodput. Three operational modes, MI (multiplicative increase), MI+ (multiplicative increase plus), and AIMD (additive increase and multiplicative decrease) are proposed in this paper, each of which takes a different strategy for adjusting the number of parallel TCP connections. We evaluate performance of the proposed automatic parameter configuration mechanism through simulation experiments. We demonstrate that the proposed automatic parameter configuration mechanism significantly improves the performance of GridFTP

VCCN: Virtual content-centric networking for realizing group-based communication

Data-centric networking has recently been getting increased attention. A representative design of data-centric networking is CCN (Content-Centric Networking), which routes packets within a network based on their content identifiers. CCN is basically designed to be open because ease of data reuse is one of the greatest advantages of data-centric networking. However, being used for real-world networking, completely open data-centric networking is not sufficient. It is required to realize closed communication within a group of users. In this paper, we propose Virtual Content-Centric Networking (VCCN), which realizes closed communication within a group of users with CCN router virtualization. This paper presents four building blocks of VCCN: extension of the content identifier, CCN router virtualization, packet transport between virtualized CCN routers, and Social Network Services cooperative user/group identification. Moreover, we implemented VCCNs basic features by extending the CCNx software and performed a preliminary performance evaluation of our VCCN implementation.

On Maximizing iSCSI Throughput Using Multiple Connections with Automatic Parallelism Tuning

In this paper, we propose an iSCSI-APT (iSCSI with automatic parallelism tuning) that maximizes iSCSI throughput in long-fat networks. In recent years, as a protocol for building SANs (Storage Area Networks), iSCSI has been attracting attention for its low cost and high compatibility with existing networking infrastructure. However, it has been known that iSCSI throughput degrades in a long-fat network. iSCSI supports a feature called multiple connections, which allows data delivery over multiple TCP connections in a single session. However, for effective utilization of the multiple connections feature, the number of multiple connections must be appropriately configured according to the network status. In this paper, we propose the iSCSI-APT that automatically adjusts the number of multiple connections according to the network status. Through experiments using our iSCSI-APT implementation, we demonstrate that iSCSI-APT operates quite effectively regardless of the network delay.

Performance Evaluation of Broadcast Communication Protocol DSCF (Directional Store-Carry-Forward) for VANETs with Two-Dimensional Road Model

In this paper, we reveal several characteristics of a broadcast communication protocol called DSCF (Directional-Store-Carry-Forward). In recent years, realization of VANETs (Vehicular Ad-hoc Networks) has been demanded for delivering information to vehicles at a low cost. In VANETs, a vehicle may or may not find other vehicles within its radio communication range because of vehicles’ mobility, which results in frequently intermittent vehicle-to-vehicle radio communications. DTN (Delay Tolerant Networking) broadcast communication mechanisms can continue communication against intermittent vehicle-to-vehicle radio communications. Our previous work, we have proposed a DTN broadcast communication mechanism called DSCF. The effectiveness of DSCF in simple one-dimensional road models has been extensively studied. However, characteristics of DSCF in a realistic environment have not been investigated. In this paper, we therefore quantitatively examine several characteristics of the broadcast communication with DSCF in two-dimensional road models through simulation. In particular, four major characteristics of a broadcast communication in VANETs are investigated: dissemination speed, uniformity, reliability, and efficiency. Through extensive simulations, we show that DSCF performs quite effectively in terms of dissemination speed, uniformity, reliability, and efficiency unless the vehicle density is too low.