Tuesday, January 27, 2015

Links of the day 27 - 01 - 2015

Today's links 27/01/2015: #NLP library, Queuing Theory, #NVM in #HPC

Monday, January 26, 2015

Links of the day 26 - 01 - 2015

Today's links 26/01/2015: #compression, text extraction, #cloud message queue comparison and #benchmarking
  • ZSTD : short for Z-Standard ( and not Zombie-STD - dirty mind...), is a new lossless compression algorithm, aiming at providing both great compression ratio andspeed for your standard compression needs. It delivers decompression a rate around ~500 MB/s per core and compressiong rate of 200 MB/s ( so no real time for you).
  • Ocropus : when you need to extract text from an image 
  • Comparison of Message Queue Architecture on AWS : the short version is that if you are not latency sensitive and don't want to spend too much effort use SQS for job management and Kinesis for event stream processing. On the contrary if latency is critical ELB or RabbitMQs (or beanstalkds) for job management and Redis for event stream processing.
  • Flotilla : Automated message queue orchestration for scaled-up benchmarking.
Compression, C├ęsar 1991

Friday, January 23, 2015

Links of the day 23 - 01 - 2015

Today's links 23/01/2015: #cloud storage - Open vStorage, #bitcoin #blockchain scalability, shortest path in corporate communication
  • Open vStorage :  its roadmap and some interesting bits on its architecture.
  • Blockchain scalability : A look at the stumbling blocks to blockchain scalability and some high-level technical solutions.
  • Surprising Facts About Shortest Paths : In a corporate communication network  the shortest path is not the fastest ( Social Network Analysis ). In other words, don’t let your train pass through the central hub for a shortcut, ’cause it’s going to stay there for a long long time.

Thursday, January 22, 2015

Links of the day 22 - 01 - 2015

Today's links 22/01/2015: NUMA & VMs,#openstack & messaging, continuous deliver

Wednesday, January 21, 2015

Links of the day 21 - 01 - 2015

Today's links 21/01/2015: cost of #bigdata scalability and when you are stuck in the middle, to big for small data and to small for big data : #mediumdata
  • Scalability! But at what COST? : Big data systems may scale well, but this can often be just because they introduce a lot of overhead. Rather than making your computation go faster, the systems introduce substantial overheads which can require large compute clusters just to bring under control. In many cases, you’d be better off running the same computation on your laptop. This has been a well know side effect in HPC as for certain type of problem high parallelism just create more overhead than accelerate the processing. The only need for such cluster is the lack of capabilities to buy hardware that is able to fit the necessary amount of data. I guess often people prefer to buy a lot of small cheaper server and say : "look i m running a cluster", than one single beefed up machine that would solve the problem faster.
  • Medium Data : when your big data is too small to warrant a cluster but too big to fit within a single machine .. I think the authors of this article and the one above should collaborate on a follow up one. Also.. one size fit all does not exist.. News at ten.

Tuesday, January 20, 2015

Links of the day 20 - 01 - 2015

Today's links 20/01/2015: Bloom filters, #Unikernel, API framework

Monday, January 19, 2015

Links of the day 19 - 01 - 2015

Today's links 19/01/2015: Data-structure books, #cloud and business agility, Leslie Lamport interview

Friday, January 16, 2015

Links of the day 16 - 01 - 2015

Today's links 16/01/2015: TCP variants, Probability books, #blockchain technology

Thursday, January 15, 2015

Links of the day 15 - 01 - 2015

Today's links 15/01/2015: #kafka, #IoT , #IBM z13

Wednesday, January 14, 2015

Links of the day 14 - 01 - 2015

Today's links 14/01/2015: HTTP , SPDY , Data virtualization

Tuesday, January 13, 2015

Links of the day 13 - 01 - 2015

Today's links 13/01/2015: #Rust Os, #virtualization Field day 4, Cpu history, Architecture design duality

Monday, January 12, 2015

Terrorism and the emergence of Stand alone complex behavior

Recently terrorist acts of seemingly disconnected individual started to rise [ 1 - 2 - 3 - 4 - 5 ]. What is rather concerning is most of those represent a phenomenon where unrelated, yet very similar actions of individuals create a seemingly concerted effort.

We are witnessing the emergence of a group of self-interested individuals who have no connections or ties to each other but who unconsciously and collectively act towards what is seen as a common purpose. This events are very quickly dismissed as isolated events and/or copycats effect. However, this a rather reductive and dismissive explanation of a new emergent demeanor.

What we have is, for the first time, a perfect combination of various groups, media outlets and information dispersal over the internet that create the prerequisite conditions for a Stand Alone Complex (STC) development within western countries. By manipulating and amplifying the fear and frustration of certain group as well as creating false information it plants the seeds for the rise of individual action loosely connected under the same banner. However, none of these event are connected in anyway beside a logo, fuzzy ideology or keywords/sentence. This is why quite often such acts are categorized under copycat / isolated individual attack.

What the authority and other entities are missing, is that without understanding and creating active countermeasures within the data stream we might very quickly face a dramatic rise of such events. This behaviors might even spread outside of the "classical" current terrorism ecosystem (religious groups/individuals tend to be more receptive but other Alter-globalization group will be quick to catch up) .

One of the most disturbing aspect of this phenomenon is that these type of individual are almost impossible to intercept before they reveal themselves through despicable actions. The natural and I would guess obvious approach for governments is to turn toward mass surveillance of population in order to establish deep profile of every individual hoping to spot sign of "auto-radicalization". And to be honest the backlash from this mass privacy invasion might just create an equivalent of a kessler syndrome within the "infosphere" with a fragmentation and radicalization of groups and individual source of content. As a result multiply the potential fertile ground for more STC individuals to rise from while making the curtail job even more difficult.

The hardest part is to understand that in order to limit the spread of stand alone complex individuals you need to first understand how the underlying fertile soil of information spreads. Second, how individuals evolve in contact with it. Third, generate counter measure to disrupt the process while also generating traps for rooting out the individuals that are too far down the road of “auto-radicalization”.  Information warfare is entering a new age, social media, bigdata, and machine learning will be the new weapons.

Links of the day 12 - 02 - 2015

Today's links 12/02/2015: a Congestion and a Scalable #network protocol, distributed #Clojure reducer, how to lie with visualizations
  • PCC : Performance-oriented Congestion Control , server side opitimization of the TCP stack trying to solve some of the current limiation.
  • Mangos : implementation in pure Go of the SP ("Scalable Protocols") protocols. An in depth overview can be found here
  • Tesser : Clojure reducers, but for parallel execution on distributed systems
  • How to Lie with Visualizations: Statistics, Causation vs Correlation, and Intuition! 

Friday, January 09, 2015

Links of the day 09 - 01 - 2015

Today's links 09/01/2015: #cloud , #storage , M&A,In Memory File system

  • WCSFSS 2014 : Workshop on Cloud Services for File Synchronisation and Sharing slides and video
  • Storage M&A : Analysis of last year and older M&A in the storage ecosystem
  • Behrooz File System : open-source distributed, scalable, replicated in-memory filesystem. BFS provides a POSIX compatible interface

Thursday, January 08, 2015

Links of the day 08 - 01 - 2015

Today's links 08/01/2015 : Slim down #Qemu in Go, #BrainFuck , Architecture analysis, Data driven #Culture

  • NoVM : lightweight type 2 hypervisor ( qemu) implemented in go 
  • Brainfuck optimization strategies : when you really want to get the best out of your brainfuck :) 
  • Panda : open-source Platform for Architecture-Neutral Dynamic Analysis using QEMU whole system emulator, and so analyses have access to all code executing in the guest and all data.
  • Data Driven : Creating a Data Culture book

Wednesday, January 07, 2015

Links of the day 07 - 01 - 2015

Today's links 07/01/2015: #google #cloud #benchmark , #IoT , #CI tales

Tuesday, January 06, 2015

Links of the day 06 - 01 - 2015

Today's links 06/10/2015: hash functions collections, monoids, functors, applicatives, and monads, category theory
  • Hash functions : collection of hash functions, ciphers, tools, libraries, and materials related to cryptography
  • Monoids, Functors, Applicatives, and Monads : the different properties of monoids, functors, applicatives, and monads packed into a neat post.
  • Introduction to Category Theory : Category theory is a mathematical approach to the study of algebraic structure that has become an important tool in theoretical computing science, particularly for semantics-based research.

Monday, January 05, 2015

Links of the day 05 - 01 - 2015

Today's links 5/1/2015: cloud object store consistency, in stream #bigdata processing and #NoSQL algos, Deterministic lockstep
  • Are We Consistent Yet? : Observed and documented eventual consistency of object stores, e.g., Amazon S3, OpenStack Swift. Your cloud datastore might not be as consistent as you think 
  • In-Stream Big Data Processing : a good insight in how to do in-stream big data processing 
  • Distributed Algorithms in NoSQL Databases : a more or less systematic description of techniques related to distributed operations in NoSQL databases. 
  • Deterministic lockstep : a method of synchronizing a system from one computer to another by sending only the inputs that control that simulation, rather than networking the state of the objects in the simulation itself