The Company

An open approach to devising cloud infrastructures

The story of Wilkushka

Wilkushka is inspired by the surname of the founder and could be translated by Wolf. It is the contraction of « wilk », wolf in Polish, suffixed in Russian with « -ушка ».

What can Wilkushka do for you?

Wilkushka is aimed at organisations which IT expertise in the cloud area is not their core business. Should it be cloud computing or dynamic systems monitoring, Wilkushka helps you devise strategies that best meet your business need in a cosy relationship at a professional level. Our commitment to you is to apply the highest standards of the industry and favour open source over proprietary solutions whenever possible.

My story

My name is Patrick Wolf. I am an independent IT consultant from Paris, France. Things have evolved since I have started my career in computing. Netscape was then the must-have browser, Alta Vista the most advanced search engine, I could browse websites developed in HTML 4.0 at the speed of light of 56 Kbit/s with my brand new V.90 modem and Linux was regarded as an operating system for spotty kids who would never defeat Windows NT.

I have been advising some of the biggest European enterprises and organisations on design and integration of open source solutions. Before starting a consulting business, I spent five years as a technical manager and architect of a Liferay web platform for a software publisher and a SaaS (Software as a Service) company for which I built an infrastructure to support multi-tenant SaaS software on AWS (Amazon Web Services). Prior to this, I was involved in the development of a network capacity planning application for the cellular network of the French Telecom company Orange and in the design and development of an embedded performance monitoring tool for the WiFi network of the French high-speed train “TGV”.

Cloud Computing

  • AWS


  • Nagios
  • SNMP
  • Wireshark
  • Snort

Operating Systems

  • Red Hat
  • CentOS
  • Debian
  • FreeBSD

HTTP and Application Server

  • Apache
  • Tomcat
  • WildFly/JBoss
  • Zope

Automation, deployment, versioning

  • Ansible
  • Git
  • Subversion

Other Skills

  • Liferay
  • BPM
  • R
  • Octave

Liferay web platform on premise and in the cloud

In order to shatter the division between development and operational teams and also provide its partners with collaboration tools and online support, the ERP editor Sage opted for a Liferay web platform which it deployed on a single physical server. The success has not kept waiting. It resulted in unacceptable performance degradation and, in the end, an unresponsive server.

After load tests, it turned out that the number of users that the platform could handle without affecting the performance was extremely low. We suggested splitting up Liferay components into a cluster of several virtualized machines made up of an Apache web server, two Tomcat servlet engines and a MySQL database. After a successful migration to a cluster and the optimisation of the JVM (Java Virtual Machine) and the database, in every use case scenarios, almost 95% of transactions remained under 2s with roughly 10,000 virtual concurrent users.

A SaaS (Software as a Service) company wanted a multi-tenant SaaS solution hosted on AWS (Amazon Web Services) to sell collaboration widgets in a Liferay web platform. The challenge was to create a model which isolates tenants from one another. We chose a model with a single VPC (Virtual Private Cloud) for all tenant deployments. The isolation happened at the level of subnets. Each tenant had its own separate application with no sharing across tenants. We succeeded in overcoming the difficulty in managing the network access control lists and security groups (AWS name for firewalls).

WiFi network performance monitoring of the French TGV train at 187 mph

Wilkushka helped the telecommunication company Orange in developing and setting up a monitoring software for the WiFi network of the French eastern train TGV. The scope of the project was on the collection of performance and QoS (Quality of Service) information of the WiFi network, the processing and transfer in the form of KPI (Key Performance Indicator) of the collected data.

We had to take into account network equipments and embedded servers, the connection between the ground and the train as well as ground servers for the end-to-end management of performance and QoS.