| KANG XINGYUAN | M, 2回目発表 | ソフトウェア設計学 | 飯田 元, 藤川 和利, 市川 昊平, 高橋 慧智 |
|
Title: Optimization of Distributed SDN Controller Placement Considering East-West Interface Overhead
Abstract: Software Defined Networking (SDN) is revolutionizing the networking industry by enabling programmability, easier administration, and faster innovation. Distributed SDN is designed to solve the problems of centralized SDN while preserving the necessary logically centralized view of the network state to simplify the development of network applications. To maintain consistency across multiple controllers, some sort of consensus algorithms are required, which may degrade the performance of the system since a consensus algorithm requires communication and synchronization between multiple nodes in a distributed environment. Moreover, these overheads will increase further if the network latency between controllers is large. Therefore, optimizing the placement of multiple distributed SDN controllers considering network latency is a challenging problem. Language: English | |||
| WEI ZHEYUAN | M, 1回目発表 | ソフトウェア設計学 | 飯田 元, 藤川 和利, 市川 昊平, 柏 祐太郎, 平尾 俊貴 |
|
Title: Segment Routing over Multipathing Quic
Abstract: In recent advancements of network protocols, the QUIC protocol has risen to prominence due to its widespread application across various practical scenarios. One of QUIC's salient features is its capability for low-latency handshakes. For instance, when a client has yet to establish communication with a specific server, QUIC necessitates only a single round-trip time (RTT) to complete the handshake process, thereby outperforming TCP+TLS in terms of speed. Moreover, for clients that have prior communication history with the server, QUIC can expedite data transfer with an instantaneous zero-RTT handshake, further diminishing latency. Despite its benefits, the application of QUIC is not without constraints. A notable limitation arises in the form of transmission congestion brought about by single line usage. Previous studies have sought to circumvent this issue through the implementation of multi-pathing in QUIC, thereby alleviating line pressure and enhancing QUIC's overall performance. Our research aims to further this improvement by incorporating Segment Routing over IPv6 (SRv6) into the existing framework. With the addition of SRv6, QUIC is afforded an expanded range of options for line selection. Thus, the integration of these two advanced technologies forms the cornerstone of our study, promising a more robust, efficient, and adaptable network protocol system. Language of the presentation: English | |||
| 鈴木 友也 | M, 1回目発表 | サイバーレジリエンス構成学 | 門林 雄基, 中島 康彦, 妙中 雄三 |
|
title:Memory-efficient RDBMS protection mechanism using location service settings
abstract:In recent years, location information has been used in a variety of applications. Since the amount of data increases over time, location information is often managed using a relational database management system (RDBMS) in a cloud environment. While RDBMSs in cloud environments have high availability and scalability, they are vulnerable to attacks from malicious cloud providers. If location information is viewed by cloud providers, privacy information such as home and workplace may be compromised. In prior research, an RDBMS protection mechanism using a hardware-level security function called Intel Software Guard Extension (Intel SGX) has been proposed, but due to the size limitation of Enclave, a secure area provided by Intel SGX, the memory efficiency is low. In this study, we propose a memory-efficient RDBMS protection mechanism that uses privacy protection settings provided in the location information service. We aim to improve the memory efficiency of the RDBMS protection mechanism by processing location information that has already been made public by the settings outside Enclave. language of the presentation:Japanese 発表題目:位置情報サービスの設定を用いたメモリ効率の良い RDBMS 保護機構 発表概要:近年、様々なアプリケーションで位置情報が利用されている。位置情報は時間経過でデータ量が増えるため、クラウド環境のリレーショナル・データベース管理システム（以下、RDBMS）を用いて管理されることが多い。クラウド環境の RDBMS は高い可用性とスケーラビリティを持つ反面、悪意あるクラウド事業者からの攻撃に対して脆弱である。位置情報がクラウド事業者に閲覧された場合、アプリケーション利用者の自宅や職場などのプライバシ情報が漏洩する恐れがある。先行研究では Intel Software Guard Extension（以下、Intel SGX）と呼ばれるハードウェアレベルのセキュリティ機能を用いた RDBMS 保護機構が提案されてきたが、Intel SGX が提供する安全領域である Enclave のサイズ制限により、メモリ効率が低いという課題があった。そこで本研究では、位置情報サービス内で提供されるプライバシ保護設定を用いた、メモリ効率の良い RDBMS 保護機構を提案する。設定によりすでに公開されている位置情報を Enclave 外で処理することで、RDBMS 保護機構のメモリ効率の改善を目指す。 | |||
| BIN BHUTTO ADIL | M, 1回目発表 | サイバーレジリエンス構成学 | 門林 雄基, 岡田 実, 妙中 雄三 |
|
Title: Aggregated Packet Pass Through Acceleration for Virtual Network I/O to Make CNFs 6G Ready
Abstract: The field of information and communication technologies is rapidly evolving to cater to the demands of emerging data and compute intensive services, requiring maximum adaptability to accommodate their ever-changing nature. Network function virtualization (NFV) is one such method to replace dedicated hardware with software network functions that can run on generic server architectures, offering network operators costs saving and flexibility. While cloud native network functions (CNFs) hold considerable promise in terms of agility, the performance of such network functions is substantially impaired by the use of virtualization and software transformation. As the development of the sixth generation of mobile systems (6G) looms and the possibility of achieving terabit-class throughput emerges, prioritizing the enhancement of CNFs has become imperative to preserve network agility and performance. Even with the state-of-the-art enhancement, the existing performance of CNFs falls short of meeting the performance demands of future 6G networks. Prior studies have identified that one of the reason behind insufficient throughput of NFV-Node is packet transferring overhead caused by virtual network I/O. While various single acceleration technologies have been explored to improve throughput, performance limitations persist. In our research, we aim to combine two orthogonal acceleration technologies zero-copy and packet aggregation to eliminate packet copy and network I/O overhead. Language of the presentation: English | |||