IT Architecture Consultant

To rationalize the release cycle of its internal applications, GroundLink decided to introduce fully isolated development, staging and pre-production environments for its development teams in Serbia. To make this transition cost-effective, they chose to leverage virtualization and dynamic provisioning.

• Profiled existing workload on physical infrastructure.
• Designed and sized the target virtual infrastructures, providing technology recommendations based on ROI evaluations.
• Coordinated deployment and migration with main office in New York.
• Trained local administrators on the new environment.
• Successfully migrated eligible projects to the new infrastructure, reducing hosting cost by 80%.

• Hyper-V
• PowerShell
• iSCSI

Infrastructure expert focusing on cross-cutting concerns, rationalization and virtualization. Conducted a total of 19 missions between 2005 and 2011.

Typical work revolved around transformations resulting from the introduction of x86 virtualization technologies in datacenters, including both technical and organizational implications.

Specific content varied a lot, and included:

• Technology intelligence, opportunity analysis, ROI estimation, IS architecture design, migration planning, audits, problem solving, development of ad-hoc tools.

➥ Next entries are selected missions.

After 2008 financial crisis, the French regulatory authority was assigned a major increase in scope and budget, leading to a ten-fold increase in IT operations department. Its organisation was deeply impacted, and in need of structuring.

• Conducted interviews to assess current state of operation processes.
• Introduced ITIL-based practices to structure operations following a continuous improvement paradigm. Most notably change management,supervision and incident management.
• Provided individual support to IT management and handled communication towards teams.
• Established weekly CAB, ensuring visibility of change decisions, with their risks and dependencies.

• ITIL
• IWS Qualities
• Nagios

To reduce IT cost of foreign branch offices, Total Petrochemicals had decided to consolidate and migrate local applications to the datacenters of the main office, creating a new virtualized infrastructure to host them.

• Inventoried branch office applications, including network flows and workload profiles.
• Designed the target architecture for both hosting platform and inter-office networking.
• Communicated with branch office teams to create procedures and tools to streamline the migration.
• Coordinated foreign branch offices and suppliers to minimize business impact of changes.
• Completed deployment by the deadline , dividing physical hosting costs by 4.

• BladeCenter
• Riverbed
• SAN
• VMware

For the launch of a consolidation plan of its 20 000-server multi-subsidiary information system, Crédit Agricole wanted to define corporate standards for virtualized hosting, based on state-of-the-art tools and practices as well as upcoming trends.

• Assessed current practices among subsidiaries, looking for synergies and pain points.
• Guided IT staff as they met major virtualisation vendors.
• Conducted workshops using gathered data and field experience as input.
• Synthesized decisions into a 2-year roadmap.

As part of its cyclical technology upgrade, LaPoste decided to assess the viability of virtualizing its 250-server ecosystem.

• Inventoried and profiled running applications.
• Designed several target scenarios with migration paths, growth plans and ROI evaluation.
• Synthesized findings and technical risks analysis in a decision document.
• Designed the detailed technical architecture after management selected a scenario.
• Transformation project achieved neutral ROI within 2 years, while dividing TCO by 6.

• VMware

Intermittent connectivity issues with foreign B2B partners led to degrading performance and occasional loss of business data.

• Set up frame capture at backbone level and designed heuristics to discriminate relevant data.
• Tracked down issue to a combination of protocol violations at equipment level and external network conditions that would compound under pathological circumstances.
• Recommended configuration changes to reduce impact of issue.
• No business data loss was reported once changes were applied.

• Cisco Catalyst
• Python
• WireShark

First deployment of a virtualized datacenter in Europe. Release-blocking performance issue in hosted applications threatened to make management abort the project. Oracle instances seemed to perform exceptionally poorly under VMware.

• Oracle internals were not well documented and Oracle refused to provide assistance.
  • Traced application behavior at system level and identified problematic patterns.
• Pioneering virtual environments meant no usable model of the workload existed at the time.
  • Isolated problematic behaviors, then instrumented them by creating test cases and benchmarks.
• Proved that virtualization was not the cause. Provided guidance on solving underlying issue.
• Prevented project from being aborted, as application went into production.

• C99
• Oracle 9i
• VMware ESX