StarfishOS: Revisiting Single System Image on CXL with State-Partitioned Microkernel

Single System Image (SSI) offers an attractive OS abstraction that enables single-machine applications to transparently execute on multiple machines with improved resource utilization. The emergence of Compute Express Link (CXL) shifts cross-machine synchronization from message passing to coherent shared memory, which may help overcome SSI's long-standing adoption barrier caused by inefficient cross-machine message passing. However, placing state in slow CXL memory degrades performance and makes the system vulnerable to partial failures, since inconsistent CXL-resident state may block or crash the whole system.