Cloud Region, safe data and environmental sustainability

Modern applications: microservices and serverless architectures 

The Cloud helps companies become data-driven by allowing them to modernise their IT infrastructure and applications with new architectural approaches such as Function as a Service, Microservices and the serverless model. It is a multi-stage journey along the IT virtualisation road with one goal: to free businesses as much as possible from the tasks and costs associated with hardware and, in part, to software, to focus on services. Services are the heart of the new customer-centric, flexible and resilient companies, capable of reacting quickly and actively to change. The public Cloud service provider takes care of the infrastructure, platform and software part; companies develop the services that add value and differentiate their business proposition in a simplified, automated way.

But what does really is a modern application? Application of modernisation starts by breaking down the (usually monolithic) legacy architecture into Microservices. Microservice architecture allows applications to be developed and organised into small independent services that communicate with each other via well-defined APIs (Application Programming Interfaces). Microservices are subsequently containerised in PaaS, e.g. executed in runtime environments that contain the bare minimum necessary to do so.

Some functions can be executed in serverless environments because they are event-driven. The event-based execution model is what we call Function-as-a-Service, or FaaS, a Cloud service that allows developers to create, execute and manage application packages as functions, without having to maintain their own infrastructure. Execution takes place in stateless containers, such as AWS Fargate, which form the basis of serverless architectures.

Serverless computing is extremely scalable and allows the provisioning of the IT resources needed to execute an event-driven workload to be fully automated. In essence, the resources remain active only while the event lasts, cutting the direct costs that would be incurred with constant availability of resources that are not fully utilised. Here, instead, the provider performs the function only for the minimum amount of time necessary for its purpose, optimising the allocation and related expenditure of resources as much as possible. If there are peaks, the allocation is increased, preventing any downtime; when resources are oversized, they are reduced. Data on these trends can be analysed to provide companies with useful information for real-time decision-making.

Database innovation: simplified storage

The serverless model can also be used for development according to the Microservices architecture and the creation of modern services and applications that fully meet the requirements of availability, reliability, performance and scalability. Scalability and development flexibility enable companies to be faster in time-to-market, experimenting quickly, correcting and validating as they go, reacting quickly to market stimuli and integrating and deploying without incurring exorbitant costs. At the same time, scalable applications built and executed with the serverless paradigm are able to autonomously adapt resources to meet peak demands, reducing both the impact on energy consumption and the effort to manage the Cloud infrastructure underpinning their operation.

But there are many use cases for the serverless model, including data management systems and ETL (extract, transform and load) services.

Public Cloud providers offer all the tools needed to modernise storage and make the most of data. AWS, for instance, offers Amazon Simple Storage Service (Amazon S3), a highly scalable object storage service that can be used for a wide range of storage solutions, including websites, mobile applications, backups and Data Lake.

AWS Lambda allows code to run without provisioning servers or managing them, thus in a serverless and event-based manner: AWS Lambda is an event-driven service and code can be configured to be started automatically by other AWS services.

As a Data Warehouse, AWS offers Amazon Redshift, a fully managed service for petabytes of data, huge volumes that more and more companies are collecting and have to manage (structured and semi-structured data generated by websites, Internet of Things, and so on). In the Data Warehouse and Data Lake, this data can be queried using standard SQL.

Finally, AWS Glue is a fully managed ETL service that simplifies the preparation and loading of data for analysis. Machine Learning as an integrated part of the database is a further element of innovation.

Machine learning: more value to your data

Machine Learning (ML) revolutionises data preparation: Machine Learning models are able to process datasets at extremely high speeds, over particularly large ranges and identifying recurring patterns, anomalies and relationships that escape the human eye. This can lead to suggestions for further, equally fast and automated analysis.

Machine Learning models can of course be trained to take into account specific interesting parameters; one example is Natural Language Processing (NLP), which is able to understand human language and improve with use (think of applications such as corporate customer service).

With its Amazon Sagemaker product, AWS allows to build, train and deploy Machine Learning models for any use case with fully managed infrastructure, tools and workflows. This allows a wider range of people in the enterprise to innovate with ML using multiple tools: integrated development environments for data scientists and code-free visual interfaces for business analysts. Structured and unstructured data are prepared on a large scale, results that used to take hours to be processed are now made available in minutes, increasing team productivity up to ten times. Sagemaker also enables over 100 billion forecasts per month and cuts cost of ownership by 54%.

Among the benefits measured by AWS in a 2020 customer survey, there is the ability to build secure and reliable ML models faster: as many ML models are used today to bring real-time forecasts to business applications and end users, it is critical to be sure that these forecasts are always available and produced quickly. Reduced costs, an abundance of different options in building ML models, compliance and support in creating datasets from data are the further advantages mentioned.

What DuneD does for its customers as AWS Partner

AWS provides a comprehensive offering of services and tools dedicated to serverless development, whereby you can outsource to AWS the provisioning and scaling of resources, maintenance and updating of operating systems and many other management tasks.

As an AWS partner, DuneD accompanies you on the path to becoming a data-driven company, including Microservices, serverless and event-driven architecture projects. We start by analysing legacy architectures and applications in order to develop a strategy for re-factoring the parts that can actually be transformed into a Microservices architecture. The ultimate goal is to make your service and product offering to your end customer more effective, reducing your development and testing costs and time-to-market.

We are specialists in Data Product, Advanced Business Intelligence (BI) and Embedded Analytics, which we implement in your company using Amazon QuickSight and expand the number of business users who can extract knowledge from data and share it through easy-to-understand dashboards.

Relying on a partner like DuneD is important not only because of the experience we bring you with our data scientists and our advice on how to conduct your IT modernisation project, but because our solutions are based on solid technological know-how like AWS, which is also reliable and secure in server and data management.

The European and Italian AWS Cloud Regions are a further guarantee in this regard. A partner like DuneD is your guarantee for a data-driven project: we can accompany you from the preparation of the data with the creation of the Data Lake to the analysis phase and the production of Actionable Insights, i.e. solid knowledge on which you can base your strategic decisions.