Cloud Migrations

In 2018, I migrated 90 servers from OnPrem to Azure IaaS, in 6 months.

Tools were deployed to improve control and cost management of the Azure estate.

In 2019, I managed the migration of c. 400 servers and 600 SQL databases into Azure (PaaS and IaaS) in 9 months and the decom of over 1,000 servers and databases.


London Stock Exchange Group


Lloyd's of London



















































MS Amlin















MS Amlin​


Jul 2020 -



Feb - Dec 

























































Nov - Dec 







Jan - Jul



Pilot project using Google Anthos and Kubernetes

I project managed the build and rollout of a new pilot project, with internal and external resources, to build a new platform that LSEG would use to pilot the containerisation of an existing website.


Migration of servers and databases off Windows / SQL 2008 onto supported versions, putting as much as possible into Azure (IaaS & PaaS) using Agile to control and manage the work allocation and delivery.

c. 400 Windows 2008 servers and 600 SQL 2008 databases migrated

c. 1,200 servers and databases decommissioned.

12 on-site staff (architects, cloud engineers, network engineer, DBA, communications, service transition, testing)

A 3rd party consultancy with 2 on-site Project Managers and 25 staff (Kiev, Bucharest)

£3.1m budget

I was the Project Manager to migrate, c. 400 Windows 2008 servers and 600 SQL 2008 databases to supported versions of Windows and SQL, moving as much as possible into Azure. This project started with this one objective, but it soon became three …

  • Migrate applications OnPrem, onto supported versions of Windows and SQL and hosted in Azure.

  • Help make the Azure environment fit-for-purpose.

  • On-board the 3rd party consultancy to enable them to analyse, develop, test and deploy the migrations.


We adopted an agile methodology using points and stories built around 2-week sprints, to calculate the work effort required, manage resources, assign work and track progress. We used Azure DevOps to allocate and track individual tasks and displays stats for Management to see progress and planned burndown.  Agile and DevOps were particularly effective for managing the offshore teams to ensure they were clear about allocated work and expectations.

The kind of tasks that I had to manage included…

  • Verify the original discovery work to improve the data integrity of the servers and databases in scope.

  • Decommission unrequired servers and databases, c. 1,200.

  • Produce design documentation and runbooks.

  • Build servers and databases in Azure and OnPrem.

    • The preference was to build new servers that met new standards and deploy the application, rather than clone existing servers to Azure (with all their legacy config and any issues)

  • Communication planning.

  • Test scripts.

  • User acceptance testing and dual running.

  • Migration of users to the new cloud-based version on an agreed date.

  • Transition to Support.

  • Decommissioning the original estate.


To help make the Azure infrastructure ready for service, the project team worked in conjunction with the Cloud Team…

  • Network and Cloud Architects and Engineers were recruited.

  • New designs and standards were created; naming conventions, standard build patterns for servers, standards for SQL servers, networking, firewalls, disaster recovery policies, backup policy, monitoring, etc.

  • Designs and standards were implemented; new load balancers and firewalls in Azure, new vNETs, new subscriptions, Terraform was used to automate the creation of new servers in Azure, Azure Backup, Azure Monitoring, etc.

  • By September, the above had been implemented and Lloyd’s Azure environment could be considered fit-for-purpose.


The 3rd party consultancy had to be setup so they could work. A mix of on-shore project managers and near-shore staff all working remotely from Kiev and Bucharest.

  • A new “role-based access control model” was devised and deployed to provide and control access to all the environments, giving the Corporation, the governance and security it required. 

  • We developed a model that could be employed through the Corporation, for production and non-prod environments, to any server or database and for any 3rd party, not just the scope of this project.

  • New servers and software were provided for the 3rd party to use; to analyse, develop, test and deploy migrations.


At the beginning, progress was slow and difficult due the delays getting the Azure environment ready and onboarding the external consultancy with the access and tools they required. Migrations began in August, discovering new issues as things were migrated and used for the very first time. 


By early September and October, output had become much faster and more efficient, getting into a ‘factory-like’ process, migrating servers and databases every week from OnPrem into Azure. 



Installing tools to improve control of VMs in the Azure cloud

Myself + one Azure engineer.

£50K budget

Once MS Amlin had been running servers in Azure for a few months, it became apparent they were missing some tools to allow them to better control and manage their cloud estate.  Having managed the prserver migration to Azure, they asked me to revisit this project and get these tools installed.

  • CloudHealth: We installed this to allow better management the Azure estate, which should result in cost savings, better understanding of VM utilisation and improving governance.

  • Log Collector by SecureWorks: They want to capture the log files from domain controllers and firewalls.

  • Access control:

  • Encryption: Additional encryption to the default storage layer was added, so servers are being encrypted twice. Now they've decided that's overkill, so it has to be removed from 180 servers.

  • Symantec Group Update Provider (GUP): is a client computer, designated to locally distribute content updates to client machines in Azure.

  • SCCM: Add a new distribution point so the global SCCM infrastructure can update servers in azure.

  • Access Control: Review all the user access into Azure an remove/limit anything that no longer or should not have access.

  • Serial Console:  Enabling this across the estate to allow low level access to each server, independent of the VMs's network or OS state.


Cloud migration from OnPrem to Azure (IaaS)

Team of 8 + stakeholders

£500K budget

6 months

I managed the migration of 91 DEV/TEST/UAT servers from a data centre in London to the Microsoft Azure cloud. 


A repeatable process was created to migrate servers to Azure.

  • Analysis was carried out on existing OnPrem servers with particular thought given to connectivity to other servers/networking that would be required when the server was moved to Azure.

  • Producing detailed migration runbooks to ensure every step was understood and documented.

  • Test scripts were created and DevOps / business users scheduled to implement testing.

  • Servers were replicated to Azure using the ASR tool.

  • Then, on migration day, the cut-over was completed, the new server in Azure tested and the original servers taken offline.

  • The project team then provided a warranty period to resolve any issues and complete the transition to BAU support.

  • The final act was to decommission the original servers OnPrem.

I ran multiple workstreams in parallel, so servers were migrated every single week until the project as completed.


© 2020 by MARCH70 Ltd

  • Black LinkedIn Icon
  • Black Twitter Icon