A change that became inevitable
About five years ago, a process aimed at redefining the organization's infrastructure scheme began in Colombia's National Natural Parks. In the past, there was an on-premise scheme of local servers; However, this scheme made it impossible to comply with all the standards that Digital Government, a public policy of the Colombian State that establishes the guidelines for the strategic use of information and communications technologies for public management, requires. For example, within the standards it is stated that the platforms of public entities must be highly available, a standard that could not be met with the on-premise platform of Natural Parks. In that sense, the challenge for the organization consisted of migrating from an on-premise infrastructure to a cloud platform that offers high availability, scalability and information security.
Regarding the people required to manage the on-premise infrastructure, the entity only had one infrastructure staff and one collaborator dedicated to servers and networks. However, there were no levels of expertise required for the management of said infrastructure. As Alan Aguia, Solutions Architect in National Natural Parks of Colombia, says, “many times the firewall or the core had to go for maintenance, which implied a downtime of all the technological services offered by the entity, reaching SLAs of 75%. at times of the year. This affected the entity since state entities have the responsibility of having information available all the time and we were not achieving it.”
In 2019, one of the big problems that Natural Parks faced was the loss of geographic information. One of the axes of the entity is associated with the production of cartography and analysis of what pressures or what changes occur in the country's coverage: Natural Parks identifies and maps the situation of forests in protected areas. This loss of information set off all the alarms. Migration to the cloud was the next step.
A migration launched
Parques Naturales chose Xertica as the partner dedicated to integrating and designing the architecture based on Google Cloud cloud technology. The challenge consisted of migrating 50 applications - which publish community services and core business applications of the entity - and implementing them in containers, which facilitates and allows the elasticity and deployment of database services.
The objective of the project was to have a solution that would allow you to easily and securely manage a CI/CD strategy to publish your services. All their Docker containerized applications were migrated to Google Kubernetes Engine and their databases to Cloud SQL. Each of the web applications is associated with a MySQL or PostgreSQL database, which is also under a high availability scheme, so that the entity is not affected by the maintenance downtime of the updates that are made on it. the system. In that sense, changes to the services were not made separately, which meant that all applications had to be unavailable when a new change was deployed in any of them.
The information persistence scheme is based on a storage scheme where the information persistence storage must be maintained over time. For each of the PoDs (a module of network components, computing, storage and applications that work together to provide network services), there were solid state hard drives in which temporary information is stored, in order to have a good performance on the applications that run in real time on the system. However, with the idea that the applications are hosted in a serverless scheme that leaves out PoDs and allows for better availability of this information, it was migrated to Cloud Functions. Front-end applications that were decoupled are now being deployed on top of Firebase, which also allows for high availability.
Finally, the StackDriver service was implemented to monitor the status of computing resources on Google Cloud Platform. Alerts and usage metrics were established at the computing level with health check for each of the servers at the CPU - Memory - Disk level.
The path to the cloud does not end
With the migration described in the previous paragraphs, an architecture orchestrated through Kubernetes Engine was consolidated, where not only elasticity and greater availability capabilities are generated, but a process for continuous integration and deployment is generated that allows changes to be made in the applications organizedly and providing the greatest possible availability.
At the organizational level, a clear benefit is to be able to integrate for the first time in the entity all the geographical information that was held in different information systems of Natural Parks. Likewise, this migration enables the possibility of capturing new sources of information from external entities and processing this data for decision making. As Alan Aguia points out, “migrating to the Google Cloud helps us simplify the time that in the past we dedicated to ad hoc technological development for the entity. We are already doing a couple of pilots with schemes in Google Cloud Dataflow to dump information into BigQuery and set up unified reporting schemes with high availability.”
