Inside our Cloud Infrastructure

Your data in the most secure and reliable data center

HEFLO is a business process management environment built on Amazon Web Services (AWS), a secure and reliable cloud services platform that delivers computing power, database storage, content distribution, and other functionalities.

AWS is the leader in cloud computing market. It is present across the globe and has several safety certifications.

SEE HOW GE USES AMAZON AWS SERVICES

Architecture

Ricardo Guimarães

CTO & Co-founder

We have designed a scalable and reliable architecture that has allowed us to grow worldwide and offer a competitive price to our customers.

Database

The data is kept in a hybrid database schema, involving Aurora DB (relational database) and DynamoDB (NoSQL database).

DynamoDB database stores tables with large volumes of data such as logs and data of process instances. The remainder data is kept in Aurora DB, which offers recovery facilities. All databases are maintained in Dublin/Ireland and São Paulo/Brazil.

We also use Redis cache technology, which maintains data in memory and maximizes application performance.

Aurora Relational Database

Replication between zones

All data is replicated in 3 different zones. This means that in the event of a data center crash, there are still 2 other available to meet the application needs.

Data Mirroring

Within each region, there are 2 copies of the data. Considering the 3 zones of replication, there will be in total 6 copies of your data!

Management by Amazon

All management is done automatically by Amazon itself. If for some reason there is a data loss or data center failure, the service is automatically redirected without loss.

SEE HOW VERIZON HAS ADOPTED AURORA DB

DynamoDB NoSQL Database

Replication of data

All data is replicated in 2 distinct regions: Ireland and São Paulo. Also, for each region there is a replication in up to 3 zones.

Geographical availability

Access and maintenance of data are always carried out from the zone closest to the user. This means shorter delays in data transfer (latency).

Management by Amazon

The maintenance between zones of availability, servers and scalability is carried out automatically by Amazon.

SEE HOW DUOLINGO USES DYNAMODB

File Repository

All files are kept in the Amazon S3 service, which was designed to deliver 99.999999999% of durability. Data is automatically distributed to a minimum of three physical facilities, which are geographically separated within an AWS region.

The folder structure isolates files by clients, and enables synchronization with a local repository by third-party tools.

Security

The HTTPS protocol encrypts all data in transit between browser and servers. The entire communication is secure.

For the relational database, there is also the encryption of data. This means that if this database falls into the hands of a hacker, the data will remain inaccessible.

SEE HOW STRIPE PROCESSES PAYMENTS ON AWS

Applications

CDN

We use CDN – Content Delivery Network – for the distribution of static content, such as the application code and website pages. Soon, we are going to apply the same technology to the distribution of dynamic content with low volatility.

Amazon has 56 distribution points. This means that there will always be a copy of the application near you, no matter where you go.

Automatic escalation

Load balancing and escalation are performed by Amazon Elastic Beanstalk technology. It configures all the rules per CPU, performance and network, as well as the default server image.

  • For every region we guarantee a minimum of two servers per service. Our servers are located in Dublin/Ireland and São Paulo/Brazil.
  • In most cases, the deployment of new versions does not cause unavailability of the application.
SEE HOW AIRBNB SCALES SERVICES ON AWS

Lambda and serverless services

AWS Lambda is a computing service that allows the execution of code without provisioning or managing servers. It runs the code only when necessary, and it automatically scales from a few requests per day to thousands per second.

At HEFLO, we use AWS Lambda to run the process engine.

SEE HOW NETFLIX USES AWS LAMBDA

Monitoring

The service monitoring is performed by the independent vendor New Relic, which provides multiple dashboards, alerts, and capabilities to identify application performance bottlenecks.

New Relic also reviews Consolidated Monitoring Information for Status Page.

Menu