In server load balancing where replicated servers are dispersed geographically and accesses from clients are distributed to replicated servers, a way of distributing the accesses from clients to an adequate server plays an important role from the viewpoint of load balancing. In the paper, we propose a new network paradigm for server load balancing using active anycast. In active anycast, an end user only sends its request to group of servers using an anycast address. When this request arrives at an active router, it selects an adequate server from the viewpoint of load balancing and changes the anycast address of a packet to the unicast address of the selected server. The decision which server is the best one from the viewpoint of server load balancing is made by an active router rather than an end user, so active anycast is a network-initiated method. Simulation results show that active anycast can accomplish efficient server load balancing, even when a small part of routers are equipped with active network technology.

A new transfer mode and a switching architecture which can support loss free and no delay jitter service class with shorter switching delay compared with "stop and go queueing scheme" is proposed. This scheme combines ATM scheme with hierarchical STM framing concept.

This paper proposes a traveling maintenance method based on the resource pool concept, as a new network maintenance model. For failure recovery, the proposed method utilizes permissible time that is ensured by shared resource pools. In the proposed method, even if a failure occurs in a communication facility, maintenance staff wait for occurrence of successive failures in other communication facilities during the permissible time instead of immediately tackling the failure. Then, the maintenance staff successively visit the communication facilities that have faulty devices and collectively repair them. Therefore, the proposed method can reduce the amount of time that the maintenance staff take for fault recovery. Furthermore, this paper provides a system design that optimizes the proposed traveling maintenance according to system requirements determined by the design philosophy of telecommunication networks. Through simulation experiments, we show the effectiveness of the proposed method.

The message level performance of error controls in data communication on ATM network is analyzed. Three layers, "a cell"(a unit of transmission), "a block"(a unit of error controls) and "a message"(a unit of transmission of user level) are considered. The error controls treated in this paper are GBN (Go-Back-N) and FEC+GBN. The cell loss process is assumed to be the two state Markov chain considering the cell loss process in ATM networks. Numerical results show that (1) the improvement of the message forwarding delay is saturated in some environments when the interface rate becomes high, (2) FEC is efficient when the burstiness of the cell loss process is small, the message length is large and the interface rate is high.

In this paper, we evaluate the performance of flow control mechanisms for reliable multicast under several retransmission approaches in terms of scalability. The mechanisms examined are a window-based flow control mechanism for ACK-based retransmission approaches and a rate-based flow control mechanism for NAK-based retransmission approaches. Our simulation results show that the NAK-based flow control mechanism has better scalability and higher throughput than the ACK-based flow control mechanism, and the delay incurred by a NAK suppression mechanism does not affect the performance of multicast flow control.

Application-level multicast has drawn a lot of attention as an alternative to IP multicast. In application-level multicast, multicast related features, such as group membership management, packet replication and packet forwarding are implemented at end-hosts instead of routers. The host perceived transmission quality and multicast forwarding responsibility depend on its position in the multicast distribution tree. This nature of application-level multicast motivates selfish members to alter their position by unrightful means to maximize their private benefits. Uncooperative behaviors of these selfish members, i.e. cheating, increase unfairness between selfish members and faithful members. In the context of bulk data distribution, this unfairness between members significantly impacts the receiver throughput. In this paper, to alleviate the negative impact of cheating members, we propose a new tree building protocol which builds dual multicast trees. Our proposed protocol constructs a shortest-widest path tree as the 1st tree. The members having lower position in the 1st tree are located at higher position in the 2nd tree in exchange for their unfairness. To investigate performance of our proposed protocol, it is compared with the existing application-level multicast protocol. Our simulation results show that our protocol outperforms the existing protocol from the view point of throughput and resource utilization against member cheating.

For recent datacenter networks, RDMA (Remote Direct Memory Access) can ease the overhead of the TCP/IP protocol suite. The RoCEv2 (RDMA over Converged Ethernet version 2) standard enables RDMA on widely deployed Ethernet technology. RoCEv2 leverages priority-based flow control (PFC) for realizing the lossless environment required by RDMA. However, PFC is well-known to have the technical weakness of head-of-line blocking. Congestion control for RDMA is a very hot research topic for datacenter networks. In this paper, we propose a novel congestion control algorithm for RoCEv2, TIDD (Timer-based Increase and Delay-based Decrease). TIDD basically combines the timer-based increase of DCQCN and delay-based decrease of TIMELY. Extensive simulation results show that TIDD satisfies the high throughput and low latency required for datacenter networks.

In the paper, we propose a congestion control scheme for reliable multicast communication which enables TCP fairness and prevents a drop-to-zero problem. The proposed congestion control scheme is rate-based one based on NAKs from receivers and cooperatively works with a flow control scheme. The congestion control scheme consists of two components of a rate-based controller and a selection mechanism of a representative. The rate-based controller runs between the sender and the representative and achieves TCP fairness and fast response to losses at the representative. The selection mechanism of the representative allows the sender to select the representative in a scalable manner, in which the sender makes use of NAKs from receivers to select it. In the paper, we also propose the switchover mechanism of the flow and congestion control schemes which enables the sender to use either of them adaptively based on network situations. When the network is congested, the congestion control scheme works to share network resources fairly with competing TCP flows. Otherwise, the flow control scheme works to adapt the transmission rate to the slowest receiver. We verify the performance of our proposed schemes by using computer simulation.

Large parts of the Internet are still incapable of native multicast, and ubiquitous deployment of multicast will take a long time. There are two approaches to provide wide-area multicast service in today's Internet. One is tunneling approach and the other is application-level multicast approach. In this paper, we focus on application-level multicast approach and propose a new scheme which improves the performance penalties of application-level multicast by making use of network support. Because in application-level multicast, endhosts provide multicast functionality instead of routers, application-level multicast delivery tree is inherently less efficient than IP multicast tree. Therefore, in our scheme, the router on the application-level multicast delivery tree alters the tree based on network-level delivery path. We evaluate our scheme with simulation experiment. Our simulation results indicate that our scheme improves the performance of application-level multicast. Further we compare our scheme to the tunneling approach from the viewpoint of transmission performances. The results reveal applicable domains of both approaches.

Content oriented network is expected to be one of the most promising approaches for resolving design concept difference between content oriented network services and location oriented architecture of current network infrastructure. There have been proposed several content oriented network architectures, but research efforts for content oriented networks have just started and technical issues to be resolved are still remained. Because of content oriented feature, content data transmitted in a network can be reused by content requests from other users. Pervasive cache is one of the most important benefits brought by the content oriented network architecture, which forms interconnected caching networks. Caching network is the hottest research area and lots of research activities have been published. This paper surveys recent research activities for caching networks in content oriented networks, with focusing on important factors which affect caching network performance, i.e. content request routing, caching decision, and replacement policy of cache. And this paper also discusses future direction of caching network researches.

CCN/NDN (Content-Centric Networking/Named-Data Networking) is one of the most promising content-oriented network architectures. In CCN/NDN, forwarding information base (FIB) might have multiple entries for a same content name prefix, which means CCN/NDN potentially supports multi-source download. When a content is obtained from multiple sources, the technical knowledge obtained for congestion control in the current Internet cannot be simply applied. This is because in the current Internet, FIB is restricted to have only one entry for each IP address prefix, which causes quite different path feature from CCN/NDN. This paper proposes a new congestion control for CCN/NDN with multi-source content retrieval. The proposed congestion control is composed of end-to-end window flow control and router assisted Interest forwarding control, and enables transmission rate regulation only on a congested branch.

In wireless environment, TCP suffers from significant performance degradation due to bit errors on wireless link and handovers because it responds to all packet losses by invoking congestion control even though packet losses are not related to congestion. Several schemes have been proposed to improve the performance degradation due to each cause. They have been evaluated in a specific network environment where either bit errors or handover occurs, i.e. they do not occur at the same time. In this paper, we reveal the packet recovery mismatch problem in an environment where both of bit errors and handover can cause the performance degradation. We pick up one scenario that TCP traffic is transmitted in the situation that ARQ (Automatic Repeat reQuest) and packet forwarding are implemented together. They are proposed to reduce the influence of bit errors and handover respectively and are natural approaches from the viewpoint of protocol layering. Computer simulation shows that in that scenario both techniques cannot perform efficiently due to interaction of each other. We also propose two buffer control approaches to resolve the packet recovery mismatch problem in our scenario according to applicability of cross-layer operation between layer 2 and layer 3.

Data center network is composed of high-speed Ethernet extended in a limited area of a data center building, so its RTT is extremely small of µsec order. In order to regulate data center network delay large part of which is queuing delay, QCN is proposed for layer 2 congestion control in IEEE 802.1Qau. QCN controls transmission rate of the sender by congestion feedback from a congested switch. QCN adopts probabilistic feedback transmission to reduce the control overhead. When the number of flows through a bottleneck link increases, some flows might receive no feedback even in congestion phase due to probabilistic feedback transmission. In this situation, queue length might be significantly fluctuated. In this paper, we propose a new delay-based congestion detection and control method. Our proposed delay-based congestion control is cooperated with the conventional QCN so as to detect and react congestion not detected by QCN.

Multicast routing problem is one of the essential problems for supporting multicast and broadcast communication service which is the most important service of the multimedia information networks. Multicast routing is the problem of finding out an adequate path which connects one source node and more than one destination node, i.e. a tree shaped path. In packet type networks, a packet for multicast communication should go through the tree shaped path as making its copies at a branching node for efficient use of network resources. However, concentration of packet copy operations at a particular node leads to performance degradation of other calls which go through this node. In this paper we propose two multicast routing algorithms which distribute packet copy operations through whole nodes in the multicast path; a link added type algorithm and a loop constructed type algorithm. Both algorithms, at first, find out an approximate solution for minimum cost path, and avoid concentration of packet copy operation at a little sacrifice of total cost in the path. Computer simulation results show that these algorithms can decrease the burden of packet copy operation per a node at the sacrifice of increase in average distance (cost) of a source-destination pair but the sacrifice of total cost is very small.

In recent years, much work has been devoted to developing protocols and architectures for supporting the growing trend of data-oriented services. One drawback of many of these proposals is the need to upgrade or replace all the routers in order for the new systems to work. Among the few systems that allow for gradual deployment is the recently-proposed Breadcrumbs technique for distributed coordination among caches in a cache network. Breadcrumbs uses information collected locally at each cache during past downloads to support in-network guiding of current requests to desired content. Specifically, during content download a series of short-term pointers, called breadcrumbs, is set up along the download path. Future requests for this content are initially routed towards the server which holds (a copy of) this content. However, if this route leads the request to a Breadcrumbs-supporting router, this router re-directs the request in the direction of the latest downloaded, using the aforementioned pointers. Thus, content requests are initially forwarded by a location ID (e.g., IP address), but encountering a breadcrumb entry can cause a shift over to content-based routing. This property enables the Breadcrumbs system to be deployed gradually, since it only enhances the existing location-based routing mechanism (i.e. IP-based routing). In this paper we evaluate the performance of a network where Breadcrumbs is only partially deployed. Our simulation results show Breadcrumbs performs poorly when sparsely deployed. However, if an overlay of Breadcrumbs-supporting routers is set-up, system performance is greatly improved. We believe that the reduced load on servers achieved with even a limited deployment of Breadcrumbs-supporting routers, combined with the flexibility of being able to deploy the system gradually, should motivate further investigation and eventual deployment of Breadcrumbs. In the paper, we also evaluate more coarse level than router level, i.e. ISP-level Breadcrumbs deployment issues. Our evaluation results show that Higher-layer first deployment approach obtains great improvement caused by Breadcrumbs redirections because of traffic aggregation in higher layer ISP.

In-network guidance to off-path cache, Breadcrumbs, has been proposed for cache network. It guides content requests to off-path cached contents by using the latest content download direction pointer, breadcrumbs. In Breadcrumbs, breadcrumb pointer is overwritten when a new content download of the corresponding content passes through a router. There is a possibility that slightly old guidance information for popular contents might lead to better cached content than the latest one. In this paper, we propose a new in-network guidance, Multiple-Breadcrumbs, which holds old breadcrumbs even with the latest breadcrumb pointer generated with a new content download. We focus on its content search capability and propose Throughput Sensitive selection that selects the content source giving the best estimated throughput. Our performance evaluation gives interesting results that our proposed Multiple Breadcrumbs with Throughput Sensitive selection improves not only throughput for popular contents but also for unpopular contents.

In large-scale data centers, two types of network are implemented: local area networks (LANs) and storage area networks (SANs). To achieve simple network management, integration of these two networks by Ethernet technology is of great interest. A SAN requires a significantly low frame loss rate. To integrate LANs and SANs, a multi-hop Ethernet configuration is generally used, and congestion may occur in traffic hot spots. Therefore, layer-2 congestion control that prevents frame loss in multi-hop Ethernet, Quantized Congestion Notification (QCN), is now discussed in IEEE 802.1Qau. In this paper, we evaluate QCN's throughput performance and reveal a technical problem with fairness among active flows. We also propose Adaptive BC_LIMIT for QCN where BC_LIMIT is adaptively decided according to current transmission rate of flows. Simulation results show that our proposed method significantly improves fairness among QCN flows.

Application-level multicast (ALM) is a feasible alternative to IP multicast. In ALM, multicast related features, such as group membership management, multicast routing and packet replication, are implemented at end-hosts instead of routers. A multicast distribution tree is constructed in the application layer, so all nodes in this tree are end-hosts. Packet transmission between end-hosts uses conventional IP unicast service. Therefore, all end-hosts can enjoy multicast communications without IP multicast service. However, ALM has a serious problem that the multicast distribution tree is intrinsically fragile and an end-host failure causes tree partitions. In this paper, to deal with this problem, we propose a new tree construction protocol which makes outdegrees of intermediate nodes be balanced. The degree-balanced distribution tree can reduce the average number of nodes decoupled by tree partitions. To investigate performance of our protocol, it is compared with an existing ALM protocol. Our simulation results show that our protocol outperforms the existing protocol from the viewpoints of robustness, loss probability and receiver-perceived delay.

In multicast congestion control, the receiver of the worst congestion level is selected as the representative and transmission rate of the sender is adjusted to TCP throughput of the representative. This approach has high scalability and TCP friendliness. However, when this approach is applied in wireless communications, wireless-caused packet loss will cause to frequent change of the representative. This is because degradation of wireless channel quality causes bursty packet loss at a corresponding receiver. Fading, one of main reasons of wireless channel degradation, is expected to be recovered after short time period, which leads to frequent change of the representative. This frequent change of representative makes the sender adjust its transmission rate to the tentative worst receiver, which brings severe performance degradation to wireless multicast. We call this technical problem, the wireless-caused representative selection fluctuation problem. Wireless-caused representative selection fluctuation problem is one of scalability problems in the wireless multicast, because this problem remarkably happens for large scale multicast. We propose two possible solutions for this problem, an end-to-end approach and a network support approach. Performance evaluation in various situation show that an end-to-end approach is sensitive for its inferring error but a network support approach shows good performance improvement.