A Case for Research Detached from the Real World
The name "Impressionism" in painting is said to have originated when a newspaper critic sarcastically titled a review "The Exhibition of the Impressionists" for a show held by Monet, Degas, and others in defiance of the official Salon of the time. Monet reportedly took a liking to the term and began to use it himself. Similarly, a research colleague once told me, "Your research is detached from the real world." Because I often make "should-be" proposals without much regard for compatibility with existing systems, I realized that this was how others perceived my work, and I actually grew fond of this critique.
My research fields are Internet architecture and distributed computing infrastructure. Devices connected to the Internet are identified by a number called an IP address. While an IP address indicates a device's location on the Internet, it is also used as its identifier or name. This becomes problematic when a device moves and connects to the Internet from a new location, as its IP address—and thus its name—changes. To solve this, I proposed separating the IP address from the name. In this proposed system, communication is enabled by specifying the device's name, which is then automatically translated to its current IP address.
Network communication protocols are defined in a layered structure, ranging from physical specifications to highly abstract rules. While increasing the independence of each layer offers many benefits, it can also decrease processing efficiency. For instance, idle states can occur because a layer cannot obtain state information from other layers. To address this, I proposed a system that allows for the exchange of state information between layers through well-defined interfaces, all while preserving the layered architecture. This enables upper layers to be aware of state changes in lower layers and prepare for the next necessary processes in advance. What had previously been handled with ad-hoc implementation workarounds for such problems, I formally organized into a protocol architecture.
The Internet's protocol stack was simple in its original design. However, as it evolved into a core part of our social infrastructure, various functions were added in a "patchwork" manner, often ignoring the layered structure. While some may argue, "If the Internet is working, it's fine as it is," I believe that continuing to add functions in this way will eventually lead to its collapse. Therefore, I proposed a new network architecture that involves organizing necessary functions, placing them in their proper layers, and adding a new layer to the current Internet protocol stack. This redesign should make it easier to add new functions in the future.
Although my work might be described as "detached from the real world," I am by no means ignoring the present state of affairs. Instead, I am looking slightly ahead at the future course of technological development. I believe this research approach is essential.