Chris Donnan : Programming – Brooklyn Style

Cool.

I recommend watching this over at Info Q for a solid overview of the space: http://www.infoq.com/presentations/Introducing-Apache-Hadoop . I like the frame here; this is a data ‘operating system’. There needs to be a macro level data rationalisation of data in the world and Hadoop is the right ‘base layer’ IMO (based on todays technology offerings out there). Even though this technology is somewhat ‘old’ at this point it is much newer than say old style relational classical SQL DB and it is now at a reasonable point of maturity for general adoption.

Google’s Spanner is worth reading about for what is now ‘newer’ tech, but not ready for mainstream consumption unless you want to build your own.

1. Setup some Hadoop for yourself – Ubuntu “quick” guide here: http://www.michael-noll.com/tutorials/running-hadoop-on-ubuntu-linux-single-node-cluster/
2. Setup Hive here: https://cwiki.apache.org/confluence/display/Hive/GettingStarted#GettingStarted-InstallationandConfiguration

I think the whole ecosystem here consisting of Hadoop, Hive here is great. Add to the list of useful related technologies that are just available now to get and use:

• Flume (get data in!)
• sqoop interop/ cope with traditional relational DBs
• Pig more ETL like tool, not sure if it is redundant with Hive/ other techs yet…
• OpenTSDB time series database – useful for capturing data that is … well – a time series (think app metric streams)

This is all makes the case for a large scale data management environment – using open source tools – that can handle massive amounts of data in many different forms.

• Ooh, Aah – Probability and Statistics Cookbook
• A few market impact function white-papers
  • Optimal Execution with Nonlinear Impact Functions and Trading-Enhanced Risk
  • Market impact and trading profile of large trading orders in stock markets
  • The market impact of large trading orders: Correlated order flow, asymmetric liquidity and efficient prices
  • Direct Estimation of Equity Market Impact
• Hidden Markov Models for Dummies

Folly can be found here

https://github.com/facebook/folly/

A few good doc refs…

Skip lists from the Concurrent Skip List ref here

Some lockless assembly fun from the Read-write spin lock here

Also includes some python examples, which I also like.

Watch it – or just get the slides

Good instructions here, but too much klunking around….

On Lion, will need to download:

• GCC 4.2 to compile with – lion uses llvm so you need to compile 4.6 with 4.2
• the new GCC itself here ftp://ftp.gnu.org/gnu/gcc/gcc-4.6.1/select gcc-4.6.1.tar.gz
• a few infrastructure libraries from here (ftp://gcc.gnu.org/pub/gcc/infrastructure/) – get mpc-0.8.1.tar.gz and mpfr-2.4.2.tar.bz2
• GMP itself needs a more recent version 5.0.2 (one of the features is 64bit under darwin)
• from the above article comments, be sure to update all your configures to include ABI=64 something like:

FastFlow (???) is a parallel programming framework for multi-core platforms based upon non-blocking lock-free/fence-free synchronization mechanisms.

Ooh, aah. This is lovely. Look out disruptor, there is nothing new under the sun apparently.

Very, very nice.A work colleague pointed FastFlow out to me yesterday and I have spent the morning reading/ playing. Big fan so far.

C++ Open Source Market Data Feed-handlers

Feed-handlers for:

• arca
• bats
• directedge
• itch
• opra

Code looks pretty well written.

Code on github – here

Fantastic Ted Talk about how algorithms are changing our world.

How Algorithms Shape Our World

C++ Optimization Manuals

Not new, but someone was asking me about this the other day… here you go – OpenOnload docs from SolarFlare

Suicide by C++ And How to Avoid It