In May, 2015, he shared a sixteen page set of PowerPoint slides, titled SILICON PHOTONICS and the DATA CENTER, reprinted in an IEEE proceedings: http://sites.ieee.org/sfbanano/files/2015/05/Fiorentino_May_2015_emb.pdf
I pulled two diagrams from that file, to share here:
The first (Slide 3 in the deck) is what THE MACHINE intends to do differently (Marco is from the Silicon Photonics side, which he whimsically titles "Almost here"):
The second slide I show is the wrap-up slide (Slide 15 of 16), after Fiorentino shows many technical illustrations:
NOW, THIS IS BIG NEWS in my view. HP Labs folk, researchers for the most part, have not been particularly adept in the past at foretelling the timeframes of 'the future' even if they have in fact invented really cool stuff.
In my post of July 7, 2014, I outlined the progression of SPECTRUM, which was HP's big bet on RISC architectures in the 1980s. A mere decade to get a high-performance machine from a set of 'known' improvements begun at HP, building on an earlier decade of work at IBM.
Hard to glean from Fiorentino's slide just where HP sits, but the first four bullets read as though there is still a little bit of invention and innovation to do to get to working equipments.
The next three--well, let's just say that the Business Case all by itself will prove daunting. I used to think that Division Managers would welcome NEW TECHNOLOGY when it is world-beating, but I found out that almost none really welcome cannibalizing everything they've worked hard to establish --HUH? JeeWhiz. Who'd have thought.... (Remember that Kodak invented Digital Photography, but the teams selling Kodak film weren't excited).
Net net--don't hold your breath just yet for the public unveiling of THE MACHINE