++AE goes to CAESAR

I have just submitted to CAESAR a lightweight parallelizable AEAD based on xor and modulo 2^b + sums called ++AE. Click on the diagram below for ++AE description.

As the diagram makes evident, only 4 sums per block + 1 cipher extra call per message are required by ++AE, plus 2 sums per associated data block.

I now that many sound designs will be submitted to CAESAR, and ++AE has now a reallly challenging process in front of it. I want to express my best wishes to all the authors competing (and specially to ++AE one ;-)

28th Feb 2014 - IOC will not go to CAESAR

After thinking over submitting IOC to CAESAR competition I have decided not to do so ... Meanwhile, a parallelizable and more robust AEAD mode has emerged applying the same principles ...

A long time ago in a galaxy far, far away from where my occupations are today, I did a Phd Thesis on Information Security in 1996 which most interesting product was an invention I called IOBC. IOBC was basically what is nowadays called an Authenticated Encryption Mode that intended to guarantee confidentiality and integrity for the data to be protected spending approximately just the same processing resources for each involved bit than the required to provide either confidentiality or integrity by the state of the art algorithms available by then.

Although IOBC may become a little bit obsoleted today by other AE modes that appeared in the field, I'm proud of this IOBC invention because, perhaps, it was one of the first AE modes with such characteristics that was ever proposed. Unfortunately, I abandoned this line of work during the same 1996 when I leaped from the academic world where I was teaching  by then to the exercise of my profession, Telecommunication Engineering, in the private sector.

With this blog I tried from 2011 to expand the limited dissemination IOBC received in its moment: it was just published as part of my Phd Thesis and presented during "IV Reunión Española Sobre Criptologia" (RESC1996) organized by University of Valladolid in 1996. For such purpose, I include bellow some links to the documentation produced then (partially translated to English from Spanish) and a cryptanalysis performed recently by Chris Mitchell from Royal Holloway / University of London.

IOBC Basics (click here)
The slides I presented in RESC1996 translated to English.

IOBC Explained (click here)

The paper that was published in RESC1996 Proceedings translated to English.

Complete analysis (click here) - in Spanish

The related Thesis Chapter in Spanish.

 

2013 Cryptanalysis by Chris Mitchell (click here)
After Chris analysis in February 2013 it comes that a forgery attack trying to modify an authentic cryptogram has a success probability around  2^-(b/3) instead 2^-(b/2) as I estimated in 1996, being b the corresponding block size of the used encryption algorithm. Although his result was a little bit disappointing for me, I really enjoyed collaborating and discussing with Chris on IOBC details. I wish to be able to come back with a "Reloaded" IOBC in some future. Chris: come prepared, next time I'll win the game ;-)

Comments and questions will be welcome (email address within the documents linked above)

Francisco Recacha ©2011, 2012, 2013.