COVER FEATURES
INTERNET OF THINGS ECOSYSTEM
The Future IoT: Building Better Legos
For smart devices to take off in any meaningful way, we need to go beyond monolithic platforms and instead create a set of secure and open protocol layers.
Scott Jenson (scott@jenson.org), Physical Web project leader, Google.
Underlying the Internet of Things: the Tizen OS
Today, the open source Tizen operating system is used in mobile devices, TV sets, cameras and motor vehicles, while tomorrow it may become an Internet of Things backbone.
Sergey Orlov (sorlov1958@yandex.ru), independent expert (Moscow).
Botnets and Internet of Things Security
Recent distributed denial-of-service attacks demonstrate the high vulnerability of Internet of Things systems and devices. Addressing this challenge will require scalable security solutions optimized for the IoT ecosystem.
Elisa Bertino (bertino@purdue.edu), professor, Purdue University; Nayeem Islam; (nayeem.islam@gmail.com), VP of research and development, Qualcomm.
A Roadmap to the Programmable World
The emergence of millions of remotely programmable devices in our surroundings will pose significant challenges for software developers. A roadmap from today’s cloud-centric, data-centric Internet of Things systems to the Programmable World highlights those challenges that haven’t received enough attention.
Antero Taivalsaari (antero.taivalsaari@nokia.com), fellow, Nokia; Tommi Mikkonen (tommi@mozilla.com), professor, University of Helsinki.
PLATFORMS
Hybrid strategy Development of the supercomputer components
Hybrid strategy of creating element base of national supercomputers provides for a combination of copying and selfdevelopment of different types of components, including processors.
Leonid Eisymont (verger-lk@yandex.ru), Expert of Scientific Research Institute FRCEC Science adviser, RSI «KVANT», Research Centre «Module» (Moscow).
MACHINE LEARNING
Deep Learning for Automated Text Processing
While shifting the paradigm in text data processing, deep learning incites disbelief in sceptical computer-savvy linguists and developers of industrial-grade data analytics solutions. Neural networks are a powerful, though quite trivial machine learning tool.
Dmitry Il’vovsky (dilvovsky@hse.ru), researcher, International smart systems and structural analysis lab; Yekaterina Chernyak (echernyak@hse.ru), professor, Higher School of Economics - National Research University (Moscow).
Machine learning vs fraud
Expert systems used to identify suspicious banks Transactions have a number of limitations. Methods of machine learning Reduce the risks associated with both erroneous detection and And fraud, and with the omission of real incidents of fraud.
Vasily Zaychenko (Vzaychenko@technoserv.com), head of the anti-fraud line of the information security competence cente, Marina Zemskova(MZemskova@technoserv.com) — database analyst, Department of Big Data, «Technoserv» (Moscow).
DBMS
Tarantool: In-Memory Database and Application Server
In-memory solutions are increasingly being used as an alternative to conventional disk-based systems that offer ‘classical’ database capabilities while providing persistent data storage. Today’s real-world projects, however, require the capabilities of both worlds.
Denis Anikin (anikin@corp.mail.ru), director of technology for email and cloud services; Sergey Pugachyov (spugachev@gmail.com), software development consultant, Mail.Ru (Moscow).
Database Benchmarks: Past, Present, and Future
Just recently, almost all significant events in the commercial database market had been illustrated with mandatory TPC-C benchmark results, while in 2017 a mere single report of this kind was published. Why the test, once so popular, lost its appeal, and how will new databases be judged hereafter?
Andrey Nikolayenko (anikolaenko@acm.org), architect, IBS (Moscow).
OPINION
Innovation as a Prerequisite for Commercial Success
Businesses that for decades had been hiring significant numbers of people to work remotely, change their ways today believing that new ideas generation is facilitated by collaborative teamwork in the office.
Michael Xie, founder, president, director of technology, Fortinet.
Smart Contracts — Dumb Idea
Smart contracts are dumb contracts, and the best contracts are fallible and human. It makes an enormous difference if we replace a legal constraint with software. The next time we consider replacing law with code to tidy up a scruffy world, let’s remember that.
Kieron O’Hara (kmo@ecs.soton.ac.uk), principal research fellow, University of Southampton.
LIBRARY
Automating Quality of Life Improvement
March, April, and May issues of the Computer magazine (IEEE Computer Society, Vol. 50, No. 3-5 2017) focus on life improvement technologies, cyber physical systems, and cutting edge learning solutions.
Alexander Tyrenko (shoorah@osp.ru), reviewer, Computerworld Russia (Moscow).
OS ACADEMY
Semantic Networks for Natural Language Processing
Semantic networks are widely used for addressing natural language processing problems. However, many of these problems are nowadays successfully solved without semantic networks thanks to machine learning. When it is enough to rely on statistical methods, and when it is more reasonable to use a high-quality semantic resource built by expert lexicographers?
Dmitry Ustalov (dau@imm.uran.ru) – Junior Fellow, Krasovskii Institute of Mathematics and Mechanics (Yekaterinburg).
