Training neural network using genetic algorithm

Over the last few years we observed the raise of artificially intelligent applications, which showed almost super-human ability to solve difficult problems. A closer look reveals that deep neural networks is what drives this wave.


On runtime code compilation

The full runtime code compilation/loading capability added to Erlang on Xen. Now there is no need to call the build service to do .beam-to-.ling transformation.


Rediscovering a cloud of the future

Today’s computing clouds, often advertised as elastic, are rather rigid. When compared to stiffness of iron, they achieve elasticity of wood. Rubber-like clouds are still on drawing boards.

Let us have a peek at the sketch of the future cloud we, the Erlang on Xen creators, have on our drawing board.


How to deploy Erlang on Xen image on Amazon EC2

It is relatively easy to build a Xen image for your Erlang application using the build service. The fast track recipe hints that the process will not take longer than 5 minutes.

Confusion ensues when the build service completes successfully and spews a ‘vmling’ image. The image requires a complete Xen toolstack to boot. While many developers have a Xen-enabled system around, this is not at all a general case.


GooFS: a ridiculously simple filesystem for cloud instances

The massive exodus to the cloud left many well-proven approaches waiting for a retirement notice. An obvious example is disk-backed filesystems.

Do we need a full-featured filesystem on every virtual node? Hardly. A virtual block device accessible by a virtual instance is much more than a handy imitation of a hard disk.


A glimpse of a truly elastic cloud

We see that migration to clouds of the large part of existing IT infrastructure is almost finished. What’s next?

From our point of view, the next wave of cloud services will use newly introduced (not inherited) features of modern platfroms. In particular, resource management in datacenters will start to avoid multi-tenancy and reduce the number of pre-started instances.


On maintaining a small memory footprint

Modern computing systems seems to encourage disregarding memory constraints. Nobody cares that a "Hello, world" program is 200K or that a Linux kernel is more than 10M in size. We have tens of gigabytes of memory after all.


Out of memory: the ultimate exception

Among unpleasant situations your application will encounter, the out-of-memory exception is probably the worst. It tends to happen deep down the entrails of the system and is notoriously difficult to recover from.