Large-Scale Systems Management

Utilizing insights from information science, such as applied probability theory, theoretical algorithms, game theory, and mechanism design, we conduct research on the design of massive data centers and network systems that highly utilize big data. Furthermore, we design services provided on these systems and study the distributed virtual currency ecosystem, represented by Bitcoin.

The foundational technology of virtual currencies, blockchain, is known to have a trilemma, unable to simultaneously satisfy the three elements of decentralization, security, and scalability. It is said to be impossible to realize a blockchain that offers high security and rapid transaction confirmation on a distributed system composed of an unspecified large number of participating nodes due to this relationship. This research theme aims to explore methodologies to overcome the blockchain trilemma through an interdisciplinary approach in information science. Specifically, it seeks to (1) mathematically elucidate the causes of the chain splitting phenomenon that leads to vulnerable security, (2) apply advanced data structures that are compressible and capable of high-speed operations to block structures and chain topology, and (3) create P2P networking technologies that enable rapid broadcast distribution of blocks. By effectively and organically integrating these technological elements, the goal is to create an extremely versatile and highly scalable blockchain technology.

The occurrence of large-scale natural and human disasters is inevitable, and thus the lifeline design is required under the assumption of large-scale disasters. In this research, we aim to establish crowd evacuation guidance using a disaster prevention nudge by designing disaster lifelines from the viewpoint of pedestrian flow, logistics, and information flow. In ordinary situations, we attempt to realize shelter allocation considering both logistics and accommodation with the help of game theory, mathematical optimization, and machine learning. In addition, we aim to establish the communication infrastructure with redundancy and resilience against large-scale disasters. On the other hand, in emergencies, we aim to establish a disaster prevention nudge assisted automatic evacuation guidance scheme, which allows evacuees to automatically collect and disseminate the disaster information. The disaster prevention nudge allows the crowd behavior to the social optimum under the individual selfish evacuation behavior.

Nowadays, various things are connected via networks. In particular, traffic on networks must be handled appropriately to ensure the safe and continuous operation of generative artificial intelligence (AI), automated driving, and/or smart factories. Network softwarization and programmability, which have attracted much attention in recent years, will accelerate the integration of machine learning (ML) and AI and realize intelligent traffic processing. In particular, AI/ML has demonstrated capabilities that rival human intelligence in certain areas. The existing in-network inference allows the AI on dedicated devices to perform advanced traffic engineering on the core network designed to operate at high throughput. However, there are still concerns about the deployment of AI in the edge network from the viewpoint of capital and operating expenditure. In this research, we aim to establish an AI-empowered network inference mechanism using SmartNICs and XDPs, which can be deployed on general-purpose devices at low cost, in order to achieve energy-efficient, lightweight, and advanced traffic processing in edge environments.

• Issues in current AI: reliability, fairness, diversity
• Vast and complex search space
• Compressed data structures such as zero-suppressed binary decision diagrams (ZDDs)
• Applications: network reliability evaluation, fair evacuation planning and political redistricting, enumerating diverse solutions