Remote working to ensure full operations to employees

The lockdown following the 2020 health emergency required the Public Administration to adapt its processes to the Remote Working mode. To minimize the operational block, in just five days an IT structure was activated which guaranteed full operations to the staff of the local PA remotely in the same ways and with the same IT tools used in traditional workplaces.

The "Comune di Cagliari" is the municipality of the city of Cagliari, the capital of Sardinia, with about 150,000 inhabitants which become about 500,000 if we consider the entire area of the Metropolitan City.
The Municipality of Cagliari employs about 5,000 people and, during the COVID-19 pandemic, decided to implement a system based on Amazon WorkSpaces to allow staff to work remotely.

The first challenge to overcome, in creating the remote working platform, was to put the Municipality in a position to be completely independent in the creation of the new VDI, starting from a series of preconfigured templates to which different roles and permissions can be applied depending on the user. In other words, the solution had to be usable by personnel without particular technical skills and without directly accessing the AWS console with extremely short VDI deployment times.
For this reason, we were asked to create an administration and monitoring console that would automate the phases of creation, stop, start, destruction of individual VDI as well as to monitor their use in terms of hours as well as view the status of other metrics. 

A second critical point in the implementation of this solution was given by the time factor: due to the urgency induced by the pandemic, in fact, the Municipality asked us to implement the system in the shortest possible time, possibly within a working week.

Failure to complete the project within the indicated time frame would have meant delaying the possibility for staff to return to
operations, blocking or slowing down the provision of services for citizens.

A solution based on Amazon Workspaces was created and integrated with an administrative and monitoring console developed ad hoc on AWS serverless technology (AWS Lambda, API Gateway, S3, CloudFront, SQS). The adoption of DevOps methodology and tools in the construction of the system has made it possible to automate all the operations of creation, destruction, activation and deactivation of the individual Workspaces, through calls to the AWS API and execution of IaC code.
Furthermore, only thanks to the adoption of DevOps methodology and tools was it possible to carry out the project in just five working days through rapid development / test / release cycles that continued later to implement new features and improve those already developed. The tools used for the realization of the project were the Infrastructure as Code code and the Serverless Framework and the AWS DevOps services (AWS Codepipeline and AWS CodeBuild) used for the release of the application, whose source code is saved and versioned in a GIT repository.

The most relevant results, compared to the challenges posed by the customer are, in summary:

1. Thanks to the DevOps approach, it was possible to go into production after a first sprint of only five days, then
   repeating further rapid development / test / release cycles to implement new features and to improve those already
   developed. It is difficult to quantitatively determine how much each delayed day would have cost the Municipality
   and Citizens, but it seems obvious that, especially at the beginning of the pandemic crisis, it was of fundamental
   importance to ensure the shortest possible interruption in the provision of services to citizens.
2. With the solution implemented, fully automated thanks to the adoption of IaC tools, we have made the customer
   independent in the creation, configuration and governance of the VDI system. The creation of a new remote
   Workstation and its release to the user, thanks to the system created, takes less than 10 '. Without this solution, the
   Municipality would have traditionally operated, distributing laptops to be configured and connected via VPN, on
   which to install the applications: the time required in this case would have been days.