{"id":4559,"date":"2018-06-14T02:38:59","date_gmt":"2018-06-14T02:38:59","guid":{"rendered":"http:\/\/filewood.snu.ac.kr\/wordpress\/?page_id=4559"},"modified":"2025-06-27T12:29:10","modified_gmt":"2025-06-27T03:29:10","slug":"research_new","status":"publish","type":"page","link":"https:\/\/redwood.snu.ac.kr\/wordpress\/index.php\/research_new\/","title":{"rendered":"Research"},"content":{"rendered":"

<\/p>\n

\"Research<\/center>
\n<\/p>\n
<\/span>1. Software-Defined Vehicles and Cloud Native Computing<\/div>
\n<\/p>\n<\/p>\n\n\n\n\n\n\n\n\n\n
\n
Developing a service-oriented software platform for software-defined vehicles based on SOAFEE and SOME\/IP <\/b><\/div>\n<\/td>\n<\/tr>\n
\n
\"\"<\/div>\n

<\/td>\n

\n
Software-defined vehicles (SDVs) have emerged as a core enabling technology for future mobility. Their ultimate success depends on the implementation of Service-Oriented Architecture (SOA). SOA is a software development\/operation methodology that separates various functions of a system into service units, connects them on a network, and builds and operates them as a system. SOA began in the enterprise computing field, and its usefulness has been proven in terms of system agility, interoperability between components and systems, deployment flexibility, and reusability.<\/div>\n
Although SOA is essential for SDVs, enterprise computing SOA technology cannot be directly applied to vehicular systems since it does not consider the functional safety of the underlying system. A software platform for SDVs must guarantee the verified performance of services along with the dynamic distribution of services.<\/div>\n
The AUTOSAR Adaptive Platform (AP) has been proposed in the automotive industry, as an industry-strength SOA platform for SDVs. While AP is currently receiving a lot of attention in the automotive industry, it has functional and non-functional limitations that have not yet been resolved as an SDV runtime platform. Additionally, AP was presented as an independent standard within the automotive industry, and thus it has shortcomings in terms of openness for cloud native computing. Consequently, open source-based software platforms for SDVs are actively underway. The related projects attempt to define architecture standards in a code-first manner based on open sources such as Kubernetes and Docker.<\/div>\n
At RTOSLab of SNU, to build the foundation for an open, cloud native SDV runtime platform, we develop a SOAFEE-based PoC (proof-of-concept) system using Kubernetes and Docker. To this end, we integrate SOME\/IP into SOAFEE and implement it on a hardware testbed that models a centralized zonal E\/E architecture.<\/div>\n
\n
<\/td>\n<\/tr>\n
\n
Enhancing SOME\/IP for real-time communications<\/b><\/div>\n<\/td>\n<\/tr>\n
\n
\"\"<\/div>\n

<\/td>\n

\n
SOME\/IP is communication middleware that supports service-oriented communications and is one of the core technologies of the software platform for SDVs. But SOME\/IP has four inherent limitations that can restrict its use in SDVs: (1) bounded message transfer latency, (2) limited throughput for huge data transfers, (3) unbounded boot time of a service instance, and (4) unbounded service discovery time.<\/div>\n
We conduct research to quantitatively measure these problems in SOME\/IP, locate bottleneck points within the protocol stack, and systematically improve them.<\/div>\n
\n
<\/td>\n<\/tr>\n
\n
System software supporting on-device AI<\/b><\/div>\n<\/td>\n<\/tr>\n
\n
\"\"<\/div>\n

<\/td>\n

\n
<\/div>\n
<\/div>\n
\n