The KIT ─ Knowledge & Information Technology
No. 133 - 1 Dec 2014
Was this forwarded to you?
In This Issue
IT Trends 2015
Most Popular ACM SkillSoft Courses
Algorithms
Internet Traffic Encryption
The Supercomputing Speed Curve
Seen Recently
Claude Baudoin

Consulting Services
  • IT Strategy
  • Enterprise Architecture Roadmap 
  • Business Process Modeling & Analysis 
  • Enterprise Software Selection 
  • IT Innovation Briefings
  • IT Due Diligence
  • Executive IT Seminars
  • Cloud Computing
  • Security Maturity
  • Software Process 
  • Knowledge Strategy
  • Technical Communities
  • Knowledge Capture
  • Taxonomy development 
  • Enterprise Social Media 
Contact Us:
c�b� IT and Knowledge Management

+1 281 460 3595
Twitter: @cbaudoin
Forward this issue to colleagues and friends: use the "forward email" link below at left, rather than "Forward" in your email software, to preserve your privacy, give the recipient more options (their own unsubscribe link, etc.) and to give us better click-through data. Thanks!

Cutter Survey on 2015 IT Trends 

Since 2006, Cutter Consortium consultants research and publish each year their forecasts of key IT trends in the new year. Will you participate this year? Take our survey and we'll send you a free copy of the most recent results, "Gaining Momentum, But Where Are We Going? IT Trends in 2014," plus $50 to spend in the Cutter Bookstore. You will also read the experts' opinions in a future issue of The Kit.

Ten Most Popular ACM Skillsoft Courses

Here is an eclectic list of online courses offered by ACM - made up of the most popular course in each of 10 different areas of specialization:
Back to Basics: Algorithms
When you learned computer science 40 years ago, there were no supercomputers and no graphical interfaces, so much of the teaching and research was about algorithms, languages and compilers, and AI. Apart from obscure writings in ACM journals, work on algorithms hasn't been as popular as in the days of Don Knuth's The Art of Computer Programming. The rapid increase in computing speeds in the following decades was probably a factor in the relative loss of interest in this discipline. A new initiative launched by the National Science Foundation (NSF), "Algorithms in the Field" (AitF) is noteworthy as it aims to foster collaboration between theoreticians and practitioners of algorithm design, verification and evaluation. Applications for grants under this program are accepted until Feb. 2015.
Internet Traffic Encryption
A year ago, in issue 109, we reported that the backlash against "domestic spying" by the US National Security Agency (NSA) had led the Internet Engineering Task Force (IETF) to promise a "fully encrypted Web" with version 2.0 of HTTP, the hypertext transfer protocol, by the end of 2014.

Well, that deadline has arrived. On the technical front, the proposed HTTP/2 standard will be submitted for consideration this month -- on time! It requires support for encryption while also offering an optional unencrypted mode. On the policy front, the Internet Architecture Board (IAB) is now calling for encryption as the default for all Internet traffic.

While encryption improves (let's not say "guarantees") privacy, and also helps protect free speech in places where it cannot be taken for granted (incidentally, there was just another report of a way to "de-anonymize" users of the Tor system), it raises a large security concern, as most malware detection tools require the ability to scan the content of internet packets. This article in the DARKReading section of InformationWeek is a good summary of the conflict (long on handwringing, short on solutions), and the thread of reader comments below the article is also very instructive.
The Supercomputing Speed Curve
The arrival of the first exascale supercomputer in the US is now predicted for 2023, according to the US Department of Energy. This means a computer capable of a sustained rate of 1 exaFLOPS, 1018 floating point operations per second. The first petaFLOPS computer (1015) came in 2008. In other words, speed doubles about every 18 months, which is faster than Moore's original law (probably because you can exploit denser circuits, but also better architectures to build faster machines).

This ComputerWorld article describes the race to exascale in US-centric terms, focusing on the risk of political funding cuts. From a global perspective, someone will get there around that time, and if not the US, then it will probably be Japan. The writer also shows his ignorance of Moore's Law, mathematics, or both, by claiming that Moore's Law should lead to an exascale machine by 2018. Exercise: find the error.
Seen Recently...

"The goal here is to blow away the reviewers via complexity. This may work for SoSyM..."

-- Excerpt of a letter received by Jordi Cabot (@softmodeling)
from someone asking him to co-author a research paper,
as reported in his blog and tweeted by Jean B�zivin.
SoSyM is Springer's Software & Systems Modeling Journal.