I am pleased to report that a short of version of the FVD-cow paper I previously posted here was just accepted to USENIX Annual Technical Conference (USENIX ATC'11), which is a prestigious and highly competitive research conference in the systems field. This shows from another angle (in addition to the good performance numbers) that FVD indeed advances the state-of-the-art in virtual machine image format. Coincidentally, Fabrice Bellard's original paper on QEMU was also published in USENIX ATC several years ago.
The draft version of the long FVD-cow paper is available at http://researcher.watson.ibm.com/researcher/files/us-ctang/FVD-cow.pdf . This paper focuses on the on-demand image streaming feature of FVD, i.e., using FVD's bitmap to efficiently implement copy-on-write, copy-on-read, and adaptive prefetching. Other features of FVD (including journal, small lookup table, and the most recent work on efficient internal snapshot) are not yet published in research paper, and will be posted here once published, which can serve as detailed documentation of the FVD algorithm and code. Regards, ChunQiang (CQ) Tang, Ph.D. Homepage: http://www.research.ibm.com/people/c/ctang
