Our Blog

Our team of specialists wanted to share some articles on technologies, services, trends and news of our industry in the era of digital transformation.

Enterprise Blockchain Use Cases

Enterprise Blockchain Use Cases

In previous articles, we went through an “Introduction to Business Blockchain” and summarized the “Technical Characteristics of Blockchain“. In this latest installment of the series, we will focus on the most widespread usage patterns and analyze a flagship case of Supply Chain, aside from Decentralized Finance (DeFi).

Use cases
When analyzing the use that is being given to Blockchain in the corporate sphere, it is possible to detect some common and recurring usage patterns:

Banking and Finance
DeFi includes digital assets, protocols, smart contracts, and Distributed Applications (dApps). It is the original use case and it involves everything related to cryptocurrencies and the financial world in general. We can mention Ripple, a global network of electronic payments, with the support of institutions such as Santander, Itaú, American Express, among others. Another example is Santander One Pay FX, a Blockchain network to streamline international transfers.

Supply Chains
After DeFi, it is the most popular use case. Blockchain allows the complete traceability of any good, from the producer to the final consumer, be it raw materials, food or medicine, in the case of Pharmacovigilance. IoT (Internet-of-Things) devices are also commonly used for automated registration in different stages of a workflow. There are already numerous success stories in several industries, such as food (Walmart and later IBM Food Trust is the most emblematic cases that we will analyze), pharmaceuticals (Novartis), automotive (Ford, BMW, Tesla), among others.

Leveraging immutability, one of the distinctive characteristics of Blockchain, stored transactions cannot be modified at a later stage, which allows a complete audit of critical information. This is used in Fraud Prevention, Claims Management, Insurance (BBVA), Health (EHR, Medical Records Management with projects such as MedicalChain & MedRec) and Pharmacovigilance (for example, the Pharmaledger project).

Public Administration
Currently, many governments around the world are conducting research on how to take advantage of the benefits of Blockchain in Citizen Identity systems, voting, budgets, and public tenders to increase the efficiency and transparency of the State. Latin America strives not to be left behind, and many developments are already underway. Argentina has several projects on the Argentine Federal Blockchain network (BFA), such as sessions of Deputies in Congress, Central Bank Complaints Registry, Citizen Files in the City of Buenos Aires, among others. Brazil, Peru, and Uruguay are also making strides. Globally, the United Arab Emirates, especially Dubai, plans to become a fully Blockchain-governed, paperless city by 2021 through the Smart Dubai project.

Certified Information
Many institutions use Blockchain for the certification and validation of all types of workforce, personal and educational records (Citizen Files in Buenos Aires).

Data Sharing
In a new human-centered model, people take control of their information, centralize it to decide who can access their data. For example, patients own their complete Medical Records and can give access to them to the professional who needs them (EHR MedicalChain & MedRec), citizens can share their credentials (Sovereign Digital Identity projects in Argentina, Data Sharing Toolkit in UAE, etc.)

Asset Tokenization
Blockchain is used for the management of digital assets, in the exchange of all kinds of goods between individuals or entities, such as tickets to shows (UEFA & Ticketmaster), loyalty program points (American Express), real estate (UK Land Registry in England), etc.

This technology is used to create records with date, time and authorship, to optimize the management of Intellectual Property. Used by Kodak with Photo Tracking and Spotify Mediachain to accurately attribute songs to creators.

Use Case: Food Supply Chain
Among the many use cases of Blockchain, its application in a supply chain is one of the most emblematic, and developments can be seen in all industries.

In a food supply chain, multiple actors have involved: farmers, ranchers, suppliers, cooperatives, packers, transporters, exporters, importers, wholesalers, retailers, and, lastly, the final consumer. Health safety is one of the biggest concerns in the food industry. Like the pharmaceutical industry, the food sector faces increased regulatory pressure from government agencies.

Walmart is a pioneer in this field, having tried several times to create a system that allows for transparency and complete traceability in the food system, which was finally accomplished in 2016. Blockchain, with its decentralized and shared ledger, seemed tailor-made for the Company’s needs. Walmart began working with its technology partner IBM on a food traceability system based on Hyperledger Fabric. For the Chinese pork industry, it allowed uploading certificates of authenticity, which gives more confidence to a system where certificates used to be a serious problem. For mangoes in the United States, the time needed to trace their origin went from 7 days… to 2 seconds!

The newly developed system allows users to know the exact origin of each item (to control disease outbreaks) in seconds, discarding only products from the affected farms. For example, it allows customers to scan a jar of baby food to see where it was made, tracing all the ingredients back to the farms.

As a result, the IBM Food Trust was launched, involving multiple companies such as Nestlé and Unilever. This platform allows access to the following information in real-time:

  • Inventory at each location.
  • The freshness of each product.
  • Average time on the shelf.

Blockchain allows a product to be traced through the different industrial, logistical, and administrative operations, from the beginning of the process to the end, and vice versa. In this way, a secure and distributed record can be consolidated with the history of every actor in the chain, their exchanges during the production and distribution processes of the product, managing information in a reliable and tamper-proof manner. As automatic transactions have no intermediaries (such as banks), they allow for faster settlements under conditions set forth in smart contracts.

IoT and Blockchain combined offer great benefits. Sensors can capture a variety of data in manufacturing facilities or transportation, transmitting all the information to a centralized repository in real-time. In turn, Managers can gain a multitude of new insights into material usage, transport conditions, etc., and apply them in planning/optimization efforts. Producers can use IoT to register the entire growth process of the product (food, pesticides, humidity, storage, location). Carriers can automatically ensure that products are moved under the right conditions of temperature, humidity, etc., thus achieving better visibility into overall logistics.

Throughout this 3-articles series, we learned about Blockchain technology and its application in the business environment. We were able to understand its basic operation, the most prominent platforms and its current application in many industries.

How can Huenei help your business with Blockchain?

  • Consultancy: We help you choose the technology that best suits your needs.
  • Architecture: Definition, deployment, and start-up.
  • Development: Smart contracts and complete systems based on Blockchain.

We work with you from the planning and definition of requirements to the start-up of the final project.

The Discovery phase in Software Development

The Discovery phase in Software Development

The two main problems that software development projects usually face are delays in delivery dates and exceeding the proposed budget; both of these issues arise when teams fail to adequately calculate the necessary resources. This results not only in a commercial failure but also in a drop in companies’ satisfaction rates.

Therefore, it is vital that companies know how to effectively conduct a digital product discovery before moving forward in laying the foundation for a project. This is why the discovery phase is crucial to mitigate the aforementioned risks, regardless of whether the project is part of a large-scale business system or if it is a turnkey development.

Both outsourcing companies and clients need to give the necessary importance to this stage, as it will improve the quality of the project and the expected final result.

What is the Discovery process?
It is the process of gathering and analyzing information about a project, its target market, audience, among other factors. It is intended to ensure a complete and deep understanding of the goals, scope, and limitations in order to help understand end-users along with their needs and requirements.

Additionally, it defines a set time to collect this information, thus allowing all members of the development team, as well as the client, to meet and create a shared understanding of the project goals. However, this goes beyond a mere kick-off meeting. This collaborative vision is about all teams being able to guide, from their point of view, more values and characteristics to ensure the execution of the project, in turn providing commercial value.

Who is involved in the Discovery phase in Software development?
Ideally, as many team members as possible should take part, from programmers and testers to functional analysts, from a more technical and specialized point of view.
On the other hand, team members from the client must also be part of it, since they have a greater knowledge of both the industry and their own consumers, offering valuable information that can increase user satisfaction rates. A list of teams and representatives involved in a Product Discovery phase would look like this:

  • Product Owner (PO).
  • Project Manager.
  • Business Analyst.
  • Solutions Architect.
  • UX Designer.
  • Programmers.
  • Quality Assurance Testers.
  • Representative End Users.
  • In theory, these should be the members involved in this phase.
    In addition to this, UX (user experience) specialists can be extremely helpful in the Discovery process, as many functional limitations can be informed by user interface (UI) requirements.

    Better integration of the design department with the Discovery processes in Software Development can be achieved through a design thinking workshop, where stakeholders meet to comprehensively discuss the project’s motivations, requirements, and vision.

    Product Discovery Process
    In order to design a Product Discovery process, the initial needs must be known. At least one initial meeting should be held with the development team so that the project can be presented to everyone, and any specific questions about the client and the project may be raised in a collaborative environment.

    Keep in mind that this process can last from a few days to a few weeks, so there is no list of steps to design the ideal process; this makes each application, idea, and team unique, and the process must be able to adapt to their needs at the time.
    However, Huenei IT Services focuses on developing the best products for your company, this is why we’ve decided to share the following steps that could help you create your Software Discovery process:

    1. Discover the purpose of software development
    This first activity focuses on the “why”.
    As we begin the digital product development phase, we ask ourselves what the ultimate goal of the project is. We can’t build a great product if we don’t know why we are building it in the first place. By understanding everyone’s expectations, we discover the motivations and context required to make focused decisions during the project execution.

    2. Get an overview of the business
    The next step is to analyze the business model of the application and understand the company behind it.
    We see this as a critical time to understand how the idea of this application came about, what the company is like, and how this product will help the company grow.

    3. Define the metrics
    The main question in this third step is how to measure the success of the product after development.
    When setting a timeline, the idea is to identify milestones and criteria to measure the product’s success.
    Goal setting methodologies are great options to deepen the discussion.

    4. Set the restrictions
    At this stage, the conversation becomes more realistic.
    Now we all know that resources are not unlimited, and that creates a scenario where each project has its limitations, such as a restricted investment or a close launch date.
    That is why we believe that it is important to know which restrictions are the most important and which allow for more flexibility.

    5. Identify the risks
    Moving on, our goal here is to identify the risks worth worrying about, so that we can focus on those that are not out of our reach.
    Just as important as listing the things that could go wrong with the project is recognizing that we can handle some risks, but not all; what these risks are and which of them could mean a leak point of resources, including time.

    6. Understand the users’ needs
    The questions in this step are related to the end-users of the application: Who are they? By discussing our ideal user and what this individual would be like, we can list how we imagine they would interact with the software.
    This process is a great instance for the team to come up with a user interface (UI) that is suitable for those users.

    7. Define work agreements and processes
    Now is the time to define how the flow of the software development process will work.
    At this stage, the team can agree on a work methodology, schedule checkpoints and other meetings, define responsibilities and work arrangements to make sure everything is clean and ready to go.

    8. Finally, build a story map
    The story mapping technique presented by Jeff Patton is a clear way to see the entire user journey that the application provides.

    In fact, knowing about Discovery in Software Development will allow you to establish a necessary approach for efficient and timely development; being clear about this from the beginning means dealing with the abundant uncertainty at the beginning of any project, that is why communication, research, and analysis are key factors in solidifying the goal and defining the direction of the product development process, as well as discovering its obstacles and risks.

    A properly planned and implemented Discovery is vital not only for the client but also for the development company in charge of the work. This not only allows teams to meet the set times and resources but also identify end-users’ needs and build a product/service that best meets their demands.
    Please visit our Software Development expertise section for more information about our development practices.

    Technical Characteristics of Blockchain

    Technical Characteristics of Blockchain

    In the previous article, “Introduction to Business Blockchain”, we learned about this technology, its main characteristics, and benefits, along with its classification and a few consolidated use cases. On this occasion, we will learn about the technical aspects that define it and briefly introduce five of the main platforms available for the implementation of Enterprise Blockchain solutions.

    The name Blockchain stems from the data structure, where a series of transactions are grouped into sequentially chained blocks. Without going into the details of a particular implementation, each block is usually made up of an index (1, 2, 3, n), a timestamp (date and time of creation), data (transactions or small programs called smart contracts), a field called a nonce (number only used once) obtained through consensus, and two hashes (fixed-width alphanumeric representation, obtained by applying a cryptographic function), one from the previous block and one pertaining to the block itself.

    The calculation of a hash is a simple mathematical operation that provides an unrepeatable result, from which it is impossible to reconstruct the source information. Thanks to the inclusion of the previous hash in the current block, any modification or elimination of a block automatically invalidates the following ones, hence blocks can only be added.

    Peer network and consensus
    Blockchain networks are always distributed and peer-to-peer (P2P). Instead of relying on a central server, the computers are directly connected to each other. Each member contributes to the computing power (consumed to create consensus) and storage. This technology is considered more secure than a centralized network, since there is no single point of attack, and it also offers 100% uptime (the network stays up as long as members are connected).

    The information sent to be added to the chain is cryptographically signed by a private key and is accompanied by the corresponding public key so that the members of the network can verify its origin. To consolidate as a new block, the parties must agree to use a consensus protocol, which ensures that the chain is the same in each node and that there are no malicious actors manipulating the data.

    Consensus protocols are mechanisms used for the members of a blockchain network to come to a consensus. In public and open networks, such as Bitcoin and Ethereum, computationally complex protocols are used, one of the best being Proof-of-work (PoW). Basically, in addition to the chain being extremely difficult to modify (it is necessary to understand in detail how it works and have control of at least 51% of the network), the purpose of the system is that anybody who intends to modify it finds it totally unfeasible to even try (it requires specialized equipment, based on devices such as GPUs or FPGAs, plus the major power use required to recalculate the blocks’ hash).

    Smart Contracts
    Smart Contracts are if … then style programs that are saved and executed by the blockchain. They get their name from legal principles, and they are intended to secure automated transactions upon meeting certain conditions in a safe way, avoiding possible malicious human actors.

    There are specific languages, such as Solidity and Vyper, and more general ones, such as Golang, Node, and Java. Support for these languages varies from platform to platform, and some don’t support them at all, usually the ones made for cryptocurrency.

    Main Enterprise Platforms
    The implementation of a blockchain at the computer program-level takes considerations regarding the security, performance, and scalability of the system to the extreme. It is highly unlikely, and not recommended if the goal is not to create a new alternative, to encode from scratch. Instead, hundreds of free platforms are available and ready to be implemented, a remarkable feature gave the security and transparency they must provide. While many are intended for cryptocurrencies, enterprise platforms also exist. A brief description of five of the most popular platforms for this purpose is included below.

    Ethereum: introduced in late 2013, and deployed for use between 2015 and 2016, it is one of the more mature alternatives. It was one of the first platforms to separate the blockchain concept from the particular case of cryptocurrencies, introducing the concept of Smart Contracts. It has its own currency, ETH, and is ideal for carrying out decentralized applications on public networks; but also, despite its lack of permits, it is widely used in business environments. It is backed by the Ethereum Enterprise Alliance (EEA, created in 2017), a non-profit organization with over 200 members, including companies among the 500 largest in the world, academic institutions, start-ups and Ethereum-based solution providers.

    Hyperledger Fabric: Hyperledger is a Linux Foundation project launched in late 2015 that brings together enterprise blockchain developments. Its best-known member is Hyperledger Fabric, initially developed, and later donated, by IBM. It is strongly permissioned and private, to the point of allowing communications between two members of the network. Due to its business focus, it uses lighter consensus protocols, allowing a greater number of operations per second. Its first version for production, 1.0, is from mid-2017, with 2.0 released in early 2020.

    Ripple: it is rooted in a pre-Bitcoin project, thus presenting some distinctive technical characteristics. It emerged as a platform in 2012, mainly for financial uses, with the main banks among its users. It is based on the use of a cryptocurrency, XRP, and unlike the main modern platforms, it does not support Smart Contracts (they are being added at the end of 2020).

    Corda: launched in 2016 by the R3 consortium, made up mainly of financial institutions. It is permissioned and has no associated currency. Its first stable version is from 2017, and it was strongly focused on banking, although over time other uses emerged.

    Quorum: a development of J.P. Morgan, announced in 2016. Basically, it is a variant of Ethereum focused on the business world, where the consensus mechanism was replaced by a faster one, and permissions were added. In the middle of 2020, it was transferred to ConsenSys, a provider of Ethereum-based business technology solutions.

    Thanks to its data structure, the type of distributed network, the cryptographic treatment, and the use of consensus, this technology boasts immutability, traceability, and security. As for the development of Smart Contracts, it requires specific knowledge about the platform where it will be executed, knowing how to interact with its API, and getting used to a new programming paradigm. Aside from the different platforms available, the business world seems mainly divided between Ethereum and Hyperledger Fabric, with the former more focused on B2C (Business to Consumer) and the latter on B2B (Business to Business).

    The main difference lies in whether the network is public and non-permissioned, or private and permissioned. To decide whether to implement one or the other, the possible need for coin consumption must be taken into account for the first case; for the second case, based on the knowledge about the members that become registered and identified entities in the network, lighter consensus protocols can be used, which increases the possible number of operations per second.

    Introduction to Business Blockchain

    Introduction to Business Blockchain

    Bitcoin, the cryptocurrency that began operating in early 2009, is the first implementation of what is known as Blockchain: a digital ledger, distributed and incorruptible, able to record all valuable information. After it became independent as a technology between 2013 and 2015, and thanks to the added execution of small programs called Smart Contracts, new use cases emerged. Around 2017, permissioned private versions, designed for the business environment, were launched. In 2020, the top implementations are at least in their second generation, with established use cases, several proposed, and many to discover.

    While the term is being widely used along with revolutionary promises in many fields, exaggerated in more than one case, it is clear that this is a technology worth taking into account and one that we should leverage. Its main benefits include immutability (impossibility of changing the recorded history), traceability, and security, allowing users to eliminate intermediaries, accelerating times, and reducing costs.


    A blockchain is a sequential data structure, replicated in a cryptographically secure peer-to-peer network made up of blocks that can only be added through consensus among its members.
    Changing the network status involves grouping data into units known as blocks. These can only be added, they include a timestamp and are mathematically validated against all previous blocks. Any attempt to delete or modify a block invalidates all subsequent ones. This mainly ensures the characteristics of immutability and traceability.

    By using a peer-to-peer network, instead of relying on a central server prone to attacks or downtime, members can communicate with each other to mathematically agree on the new state of the network. This mechanism and the intensive use of cryptographic techniques provide an unprecedented degree of IT security. Along with an ability to break the cryptography and control at least 51% of the network members, and alteration attempt would require a computing power capable of reconstructing the new state of the network, all of which would render the task unfeasible and inconvenient.

    Blockchain networks can be public or private, with or without permissions, thus creating the following main categories:

    • Public, not permissioned: anyone can participate and all the information is openly available. It is the preferred option for the implementation of cryptocurrencies and distributed applications (where transparency is provided and censorship is avoided). These networks use greater computational complexity, which makes them slower. There is no owner.
    • Public, permissioned: each participant’s identity is verified to grant access, but all information is openly available to participants. It is usually used in voting systems and is implemented by a consortium of public and private actors.
    • Private, permissioned: each participant is verified and approved. It is usually the preferred business option. The owner of the implementation is a company or a small group of companies. Mathematical complexity is decreased based on the reduced number of participants and their permissions, so the network is much faster. Some reach the granularity of creating specific channels between two participants, without affecting the mathematical security provided by the network.

    A blockchain is at the backend of an application, taking the place of a traditional database. Since its storage capacity only increases, and it is also replicated in all the members of the network, it is recommended to keep only what is strictly necessary, using a database for everything else.

    You don’t need a blockchain per se, but you need the solution to a problem that could eventually lead to its implementation.

    As two key points to decide on its implementation, users should analyze whether there will be different actors first, and if so, if more than one will be granted writing permission. If both conditions are not met, there are simpler solutions. A good reason for implementation is the need to eliminate intermediaries, allowing two or more parties to interact relying on the blockchain.

    If trust is not an issue, the use of blockchain does not represent any advantage over a database.

    Use cases

    • It is possible to detect some common usage patterns, including:
    • Decentralized Finance (DeFi), through a secure platform for making optimized payments, executing insurance, etc.
    • Digital asset management, for the exchange of all kinds of goods between individuals or entities, such as tickets to shows, loyalty program points, real estate, etc.
    • Supply chains, for the complete traceability of any product, from the producer to the final consumer, whether it be raw materials, food, or medicines, in the case of Pharmacovigilance. IoT (Internet-of-Things) devices are also commonly used for automated registration in different stages of a workflow.
    • Governance, with the implementation of Citizen Identity systems, elections, budgets and public tenders to increase the efficiency, quality, and transparency of the State.
    • Management of Intellectual Property, creating a record of the date, time, and authorship as reliable evidence certifying this data.
    • Certified information of all kinds, ensuring its veracity, for example, University degrees.
      Secure data sharing, centralizing information and allowing access to it to those who need it, for example, medical records or research data.

    As a concept, blockchain is relatively new, with its first implementation (Bitcoin) in 2009, its opening to other uses in 2015, business versions emerging in 2017 and multiple proofs of concept in 2020. However, it involves the combination of multiple technologies that have been around for a long time and were creatively combined in a platform with disruptive uses.

    Some consider that hearing the word “Blockchain” today is like talking about the Internet in the mid-90s, and we are reminded about the way the Internet transformed the world we live in for business, commerce, communications, and media. In fact, public blockchain networks are often compared to the Internet, while private networks would be more of a synonym for Intranet.

    Without having to be certain about the future, it is clear that it is a tool worth knowing, mastering, and using. Not only does it add value and security in well-known use cases, but it also opens the door to new opportunities. Furthermore, in our currently globalized world, it is a matter of time before having to join an existing network will require the use of this technology. Huenei can help your business by providing Consulting, Network Architecture Design, and Application Development services, from the planning and definition of requirements to the final project deployment phase.

    Good practices on DevOps

    Good practices on DevOps

    In our previous article, “Key points about DevOps” we talked about what DevOps is and how it can help you optimize key elements in your company, such as shortening launch times, accelerating the realization of new products and services, and reducing costs. In this article, we will learn more about some good practices to start implementing this as soon as possible.

    Doing a brief review, DevOps refers to the words ‘Development’ and ‘Operations’, it is a discipline that, despite being relatively recent, has allowed many companies and organizations to rethink their processes for more efficient and agile ones, managing to increase its competitiveness and efficiency.

    Good practices on DevOps
    For sure, there are countless benefits to the implementation of this methodology, taking into account that it does not imply an improvement in terms of technology or productivity in itself, but rather it allows to streamline the level of communication and collaboration between departments for optimal execution of operations, time and quality of delivery.

    Adopting this new process does not happen overnight, and expected results could be negatively impacted if the company implements ineffectively.

    At Huenei, we use this methodology to increase the effectiveness of our software development teams, as well as to improve the quality of the constant deliveries in Custom Software, Mobile Applications, involving it in Testing & QA and UX / UI processes. So we recommend following these practices:

    1 – Continuous Integration Process (CPI)
    This process is strongly supported by agile methodologies, its principle is to reduce implementation and deployment times, managing to divide a project into more parts, making gradual deliveries.

    How this could help? As the team of developers constantly review changes in the repository code, they can more quickly detect any type of flaw and even constantly improve the execution of the operation. It is these early findings in the software life cycle that will help the development department to solve problems almost on the spot (and even prevent them).

    2 – Continuous delivery
    Arguably, this step accompanies the previous one, as starting with development, build, unit testing, static code analysis, and static analysis security testing in the PCI, it promotes automation of functional tests, integration, performance, and security, along with configuration management and deployment.

    As the latter are critical areas for automation, and essential in a DevOps context, it is a type of practice that increases the amount of testing and verification at various stages in manual and automated code cycles.

    In addition to allowing a team to build, test and launch the codebase in a faster and more frequent way, which by dividing it into cycles of smaller size and duration, speeds up the processes of organizations, allowing them to carry out more launches, reduce manual deployments and reduce the risks of production failures.

    3 – Communication management
    A key element in DevOps management, since its focus is on keeping all interested parties related to development, operation, and implementation informed, taking into account that we are integrating the tasks of different departments. Therefore, communication is essential and becomes a fundamental element for a total adoption, keeping everyone on the same page to involve them in the whole process, keep up to date with the organization and avoid doubts between them or the final products.

    To apply the strategy correctly, it is vital to keep all the teams and members up to date, in this way it is guaranteed that the leaders of the organization (from the sales, production, and management departments) can be involved in the processes and guide the team development to make successful changes, from their perspectives and knowledge.

    4 – Test automation
    In software development, regular testing is essential to create quality code, which is vital to implement in a DevOps management, not only because it saves time by quickly completing tedious and time-consuming tasks, but also because it allows you to identify quickly early failures in pre-deployment processes.

    This in itself is the core of agility and innovation, since the additional time that team members will have can be invested in tasks with greater added value or in monitoring the results of new processes, identifying failures and opportunities for improvement.

    5 – Continuous monitoring
    Like all cultural and work process changes, this requires close and continuous monitoring to anticipate and identify that all actions are being carried out correctly and that performance is as expected.

    Continuous delivery of feedback from all members of the organization in real-time is vital to the organization; from the production team that runs the application to the end customer at the official launch stage. In this way, we ensure that the developer can benefit from all the valuable feedback on the end-user experience in the process and in turn, modify the code to meet the expectations of the end-user.

    With an ever-evolving business environment and an ever-improving and growing technology landscape, organizations must seek to stay ahead, with DevOps they can increase the speed and quality of software implementations by improving communication and collaboration between parties. interested.

    In themselves, these good practices allow creating a clear guide to guide any company, of any size and industry, to immerse themselves in the necessary cultural change, managing to increase productivity and efficiency through high-quality deliveries, adding transparency to their processes. and open collaboration across development and operations teams.

    Know more about our process in the Software Development section.

    Hardware-accelerated frameworks and libraries with FPGAs

    Hardware-accelerated frameworks and libraries with FPGAs

    In the previous articles, we talked about Hardware Acceleration with FPGAs, the Key concepts about acceleration with FPGA that they provide, and the Hardware acceleration applications with FPGAs. In this latest installment of the series, we will focus on Hardware Accelerated Libraries and Frameworks with FPGAs, which implies zero changes to the code of an application. We will review the different alternatives available, for Machine and Deep Learning applications, Image, and Video Processing, as well as Databases.

    Options for development with FPGAs
    Historically, working with FPGAs has always been associated with the need for a Hardware developer, mainly Electronic Engineers, and the use of tools and Hardware Description Languages (HDL), such as VHDL and Verilog (of the concurrent type in instead of sequential), very different from those used in the field of Software development. In recent years, a new type of application has appeared, acceleration in data centers, which aims to reduce the gap between the Hardware and Software domains, for the cases of computationally demanding algorithms, with the processing of large volumes of data.

    Applying levels of abstraction, replacing the typical HDL with a subset of C / C ++ combined with OpenCL, took the development to a more familiar environment for a Software developer. Thus, basic blocks (primitives) are provided, for Mathematical, Statistical, Linear Algebra, and Digital Signal Processing (DSP) applications. However, this alternative still requires a deep knowledge of the hardware involved, to achieve significant accelerations and higher performance.

    Secondly, there are accelerated libraries of specific domains, for solutions in Finance, Databases, Image, and Video Processing, Data Compression, Security, etc. They are of the plug-and-play type and can be invoked directly with an API from our applications, written in C / C ++ or Python, requiring the replacement of “common” libraries with accelerated versions.

    Finally, we will describe the main ones in this article, there are open source libraries and frameworks, which were accelerated by third parties. This allows us, generally running one or more Docker instances (on-premise or in the cloud), to accelerate Machine Learning applications, Image processing, and Databases, among others, without the need to change the code of our application.

    Machine learning
    Without a doubt, one of the most disruptive technological advances in recent years has been Machine Learning. Hardware acceleration brings many benefits, due to the high level of parallelism and the enormous number of matrix operations required. They are seen both in the training phase of the model (reducing times from days to hours or minutes) and in the inference phase, enabling real-time applications.

    Here is a small list of the accelerated options available:

    TensorFlow is a platform for building and training neural networks, using graphs. Created by Google, it is one of the leading Deep Learning frameworks.

    Keras is a high-level API for neural networks written in Python. It works alone or as an interface to frameworks such as TensorFlow (with whom it is usually used) or Theano. It was developed to facilitate a quick experimentation process, it provides a very smooth learning curve.

    PyTorch is a Python library designed to perform numerical calculations via tension programming. Mainly focused on the development of neural networks.

    Deep Learning Framework noted for its scalability, modularity and high-speed data processing.

    Scikit-learn is a library for math, science, and engineering. Includes modules for statistics, optimization, integrals, linear algebra, signal and image processing, and much more. Rely on Numpy, for fast handling of N-dimensional matrices.

    XGBoost (Extreme Gradient Boosting), is one of the most used ML libraries, very efficient, flexible and portable.

    Spark MLlib is Apache Spark’s ML library, with scaled and parallelized algorithms, taking advantage of the power of Spark. It includes the most common ML algorithms: Classification, Regression, Clustering, Collaborative Filters, Dimension Reduction, Decision Trees, and Recommendation. It can batch and stream. It also allows you to build, evaluate, and tune ML Pipelines.

    Image and Video Processing
    Image and Video Processing is another of the areas most benefited from hardware acceleration, making it possible to work in real-time on tasks such as video transcoding, live streaming, and image processing. Combined with Deep Learning, it is widely used in applications such as medical diagnostics, facial recognition, autonomous vehicles, smart stores, etc.

    The most important library for Computer Vision and Image and Video Processing is OpenCV, open source, with more than 2500 functions available. There is an accelerated version of its main methods, adding more version after version.

    For Video Processing, in tasks such as Transcoding, Encoding, Decoding and filtering, FFmpeg is one of the most used tools. There are accelerated plugins, for example for decoding and encoding H.264 and other formats. In addition, it supports the development of its own accelerated plugins.

    Databases and analytics
    Databases and Analytics receive increasingly complex workloads, mainly due to advances in Machine Learning, which forces an evolution of the Data Center. Hardware acceleration provides solutions to computing (for example with database engines that work at least 3 times faster) and storage (via SSD disks that incorporate FPGAs between their circuits, with direct access to data processing). the data). Some of the Accelerated Databases, or in the process of being so, mainly Open Source both SQL and NoSQL, are PostgreSQL, Mysql, Cassandra, and MongoDB.

    In these cases, generally what is accelerated are the more complex low-level algorithms, such as data compression, compaction, aspects related to networking, storage, and integration with the storage medium. The accelerations reported are in the order of 3 to 10 times faster, which compared to improvements of up to 1500 times in ML algorithms may seem little, but they are very important for the reduction of costs associated with the data center.

    Throughout this series of 4 articles, we learned what a device-level FPGA is, how acceleration is achieved when we are in the presence of a possible case that takes advantage of them (computationally complex algorithms, with large volumes of data). General cases of your application and particular solutions, ready to use without code changes.

    How can Huenei help your business with Hardware Acceleration with FPGAs?

    Infrastructure: Definition, acquisition and start-up (Cloud & On-promise).

    Consulting: Consulting and deployment of available frameworks, to obtain acceleration without changes in the code.

    Development: Adaptation of existing software through the use of accelerated libraries, to increase its performance.