seems like a hack. the article is more of a puzzle than journalism, though: where do the pieces fit? FC-type SANs are expensive and complicated. any server can boot over PXE, but a certain common OS seems not to like being diskless. iscsi could solve this disparity, since putting blocks onto a network interface can be done cheaply, efficiently and with the hardware that every server already has - alas SAN thinking seems to have spoiled that field. esata is puzzling though, since most servers don't even bring sata out of the chassis - even using USB would probably make more sense (not as fast or efficient, but remember, we're only talking about booting.)
I'm not sure what you find confusing. Is there a question?
BTW, it turns out LSI sees the Syncro CS as the bigger opportunity because it will ride what Microsoft hopes is a big trend to cluster-in-a-box OEM/ODM systems that act as powerful but simple SMB and branch office systems with multiple servers using shared JBOD storage.