<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Mario Camarca on Medium]]></title>
        <description><![CDATA[Stories by Mario Camarca on Medium]]></description>
        <link>https://medium.com/@mario.camarca?source=rss-201c9e8a14a6------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*up2uH5ZaNw-ja6VKFX5AGw.png</url>
            <title>Stories by Mario Camarca on Medium</title>
            <link>https://medium.com/@mario.camarca?source=rss-201c9e8a14a6------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 07:03:47 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mario.camarca/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[My journey in Cloud Engineering: documenting the third module [System Design Applications].]]></title>
            <link>https://medium.com/@mario.camarca/my-journey-in-cloud-engineering-documenting-the-third-module-system-design-applications-cd19c0fbcb67?source=rss-201c9e8a14a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd19c0fbcb67</guid>
            <dc:creator><![CDATA[Mario Camarca]]></dc:creator>
            <pubDate>Sun, 08 Dec 2024 17:37:20 GMT</pubDate>
            <atom:updated>2024-12-08T17:37:20.773Z</atom:updated>
            <content:encoded><![CDATA[<p>In this module, I will compare and contrast several cloud architecture designs: monolithic vs microservices vs serverless. After that, I will show also the assigned projects involving architecture design!</p><p>A little “disclaimer” before proceeding: as already stated before, I want to remark that in these articles I will not copy/report any part of Cloud Engineering Academy study material (which off course is protected by copyright). What you can find, instead, is my own personal summary.</p><p><strong>Monolithic Architecture.</strong></p><p>Monolithic architecture refers to a software development model where a single codebase executes multiple business functions. This “monolithic”, single, non-divisible unit, is composed by:</p><p>- User interface/frontend</p><p>- Business logic/backend</p><p>- Data interface/data access layer which will communicate with a database</p><p>The 3 tiers are encapsulated together, and the app will be deployed as a single jar/war file. See below a graphical representation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/451/1*l_oCH52wKFmIM9saGuydhg.png" /></figure><p>For Monolithic we can summarize the following pros and cons.</p><p>Pros: simplicity and uniformity as the application is easier to develop, test, and deploy.</p><p>Cons: flexibility, scalability. This because it is hard to implement new features, any change we make would require deploying a new version of the entire application.</p><p>Why we still consider this monolithic architecture? Because this approach is still suitable for smaller programs that require quick and cheap deployment while providing integration with legacy systems.</p><p>To recap: in Monolithic architecture the application tiers are encapsulated in a single unit. As we already know (see my previous article), a traditional application has a frontend part and a backend part (API, logic, services, load balancer, and database). As an example, let’s imagine to build everything together and deploy it on a VM (which will be our server), that would be a Monolithic architecture design, as the services tightly couple together.</p><p><strong>Microservices Architecture.</strong></p><p>Microservices Architecture is a software design pattern that structures an application as a collection of loosely coupled services. Each service is independent and provides different business functions. The services communicate with each other through lightweight protocols, such as APIs and each service communicate with his own database.</p><p>Given that each microservice is and independent unit corresponding to a single business capability, any change to one service does not impact the other.</p><p>Here below a sketch representing microservices design.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/776/1*45VZnHMD5TN5S6QvuqtwDQ.png" /></figure><p>Before going further, I want to quickly introduce the concepts of API and API gateway, as these are important concepts to understand for what will be showed later on.</p><p>- API (Application Programming Interface): it’s a set of defined rules and protocols that enables different software programs or systems to communicate with each other, exchanging data and functionality, allowing them to interact and integrate seamlessly. An API also can provide a list of operations that the developers use to interact with a service of app. APIs are important cause they enable integration/communication of services, and automation of tasks.</p><p>- API Gateway: it is a management tool that sits between clients and backend services, acting as an entry point for API calls, enforcing security, scalability, and high availability, and providing features like traffic management, and authentication.</p><p>So, how a Load Balancer compares to an API Gateway? In a modern Microservices architecture, Load Balancers and API Gateways are two essential components that manage network traffic and ensure scalability, security, and reliability. While they share some similarities, they serve distinct purposes and are not interchangeable. Load balancers focus on distributing traffic, granting availability and performance, while an API gateway does more as it makes sure interactions within protocols, and that the app is secure and efficient.</p><p>See below a graphical representation showing the API Gateway role in Microservices.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gPia1HxdytH8KJdF3fUBvQ.png" /></figure><p>To recap, by performing all the aforementioned functions, an API Gateway simplifies the complexity of a Microservices architecture, providing a centralized point of control and management for client requests, and optimizing resilience, scalability, and security.</p><p>Before ending this paragraph, let’s also recap the main pros and cons of Microservices.</p><p>Pros.</p><p>Scalability: each service can be scaled individually without affecting the entire application, allowing the business to easily adapt to changing demands.</p><p>Flexibility: independent services allow for easier maintenance and evolution of individual components.</p><p>Resilience: If one service experiences issues, others can continue to function, reducing overall system downtime.</p><p>Cons.</p><p>Complexity: this design carries additional complexity due to the need for service discovery, communication, and orchestration.</p><p>Higher operational overhead: managing multiple services requires more infrastructure, monitoring, and logging, increasing operational overhead.</p><p>Security and governance: microservices architecture requires additional security measures and governance structures to ensure proper access control, authentication, and authorization.</p><p><strong>Serverless Architecture.</strong></p><p>Serverless does not mean there is no server. In this model, the applications are built without managing servers or infrastructure. That means cloud providers manage and scale resources automatically, allowing developers to focus solely on writing code.</p><p>Let’s see at once the main pros and cons of this approach.</p><p>Pros.</p><p>No server management: as mentioned, cloud providers handle all the infrastructure responsibility (server provisioning, scaling, and maintenance).</p><p>Scalability: Cloud providers automatically scale resources to match demand, eliminating the need for manual scaling.</p><p>Cost-effective: Developers only pay for the compute time consumed by their code, reducing costs and eliminating idle server expenses.</p><p>Improved developer productivity: The developers’ focus is on writing code without worrying on managing infrastructure.</p><p>Cons.</p><p>Limited control: Developers have limited control over the underlying infrastructure and scaling</p><p>Complexity: Debugging serverless applications can be more challenging due to the ephemeral nature of functions.</p><p>Popular serverless platforms are: AWS Lambda (Amazon Web Services), Azure Functions (Microsoft Azure), Google Cloud Functions (Google Cloud Platform), Cloudflare Workers (Cloudflare).</p><p>See below a sketch representing Serverless architecture.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/807/1*Grsr7kC_kq3EfDHf-SCiLw.png" /></figure><p>On top we have “User applications”, this represents any user application (desktop, mobile…) that can make HTTP requests to an API. The request will go through the API Gateway that will route it to the appropriate serverless function. From the representation one may think this architecture is very similar to Microservices, so let’s examine the differences.</p><p>While Serverless is similar to Microservices in terms of structure, the functions related to each service are not always running as the resources are scaled, by the cloud provider, based on the demand. In other words, on Microservices the functions related to the services are always ON and running, whereas on Serverless they are running but not active. Both support developing scalable and flexible applications, but they differ on how they manage the cloud infrastructure and services resources. See below a summary of the key differences for each point.</p><p>· Server management: Serverless eliminates server management, while Microservices require internal or external teams to manage and maintain individual services.</p><p>· Function state: Serverless functions are stateless, whereas Microservices can maintain state (retains information about the current state of a user’s interactions) and have their own databases.</p><p>· Scalability: Serverless scaling is automated by the cloud provider, while Microservices requires manual scaling or orchestration (for example you will need Kubernetes).</p><p>· Cost: Serverless pricing is based on usage, while Microservices requires upfront investment in infrastructure and team resources.</p><p>· Complexity: Microservices introduces additional complexity due to service interactions, while Serverless functions are simpler and more focused.</p><p><strong>Project — Simple e-commerce application.</strong></p><p>The aim to this project is to get practical insights into the advantages, challenges, and use cases of each architectural style explored in this article.</p><p>For this project, I will start with Monolithic architecture, then refactor it into Microservices, and finally implement a Serverless approach for certain functionalities.</p><p>All right, let’s pretend a small company gave a task to design a simple e-commerce application.</p><p>So, for this use case a Monolithic design seems to fit well as we expect a small amount of concurrent users for our e-commerce app. Also, for a small business, this is a very good cost-effective solution.</p><p>See below the monolithic approach for the e-commerce app.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/224/1*-Cyux2HDYULl4hQ5v5vmqg.png" /></figure><p>The core functions of the application are: “user registration” and “product catalog”, which will be the front-end part of our application, where there will be our static content such as the buttons to register and products to browse. Then we have the “shopping cart” and “payment processing” which are part of the backend as these will be the logic part to handle user’s actions for buying products. Finally, we have a database containing the data, such as users table, products/prices.</p><p>Monolithic architecture serves well a small or medium-sized business due to the simplicity, uniformity, and its cost-effectiveness as the single codebase structure requires less infrastructure overhead, it is easier to manage, deploy, and maintain.</p><p>However, although the deployment can be simpler, the fact that a Monolithic application requires deploying the entire codebase as a single unit can result in longer deployment times for the company, as all the development activities (design, packing, testing, deploying) require to be done all together for each component. Another disadvantage is related to limited flexibility since making changes will require modifying multiple parts of the codebase. For example, if our business requires some changes relating the “shopping cart” service, we will have to re-deploy the entire application.</p><p>At this point, let’s suppose that the business is growing at very fast pace, we have more resources, and so now we really need high scalability and flexibility to sustain the growth. Therefore, we refactor the app into Microservices.</p><p>See below the Microservices version of the e-commerce application.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*uGUJQgT30mffhW0514sFBw.png" /></figure><p>In a Microservices design it is very important an API gateway. As we saw earlier, this acts es the entry point for all client requests and then will route each to the appropriate microservice. Each of those is responsible for a single functionality or feature and, unlike the Monolithic design, they can be developed, deployed, and scaled independently. Instead of sharing a single database, each microservice communicates with his own database. Here below is a description of each microservice function:</p><p>- User Service: to handle user registration/authentication</p><p>- Product Service: to display the product list where the users can choose from</p><p>- Cart Service: this will be the shopping cart where users will be able to add products and place an order</p><p>- Payment Service: this will handle the payment processing</p><p>With Microservices, now each component of the application can be scaled independently. For our e-commerce app, that means it is possible to scale only the parts that need to handle more traffic, rather than scaling the entire application. We can also add another microservice and integrate it with the pre-existing application without having to rewrite the entire app. Microservices brings also advantages in terms of flexibility as we can use different technologies and programming languages for different services based on their specific requirements, without being limited to a single technology stack.</p><p>However, there are also disadvantages since this design can introduce more complexity requiring a robust deployment pipeline and automated tools to ensure that updates are deployed smoothly, and without downtime. Since each service is independent, with Microservices we face more challenges to monitor, debug (tracing issues among multiple services can be more complex), and test the application. There are also challenges for the management and orchestration as Microservices require specialized tools and expertise, such as Kubernetes or Docker Swarm, to ensure smooth operation and scalability.</p><p>How we can address these disadvantages we are facing with Microservices? We can have various serverless Lambda functions to handle the tasks that before were assigned to Microservices components. See below the architecture diagram updated with Serverless functions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/780/1*Q5Vd3AMMPGWElX2eg1yZOQ.png" /></figure><p>Doing so will bring advantages in terms of scalability and management as we don’t have to worry about provisioning or managing servers. Implementing Serverless will also bring cost-effectiveness since our business will only pay for the compute time consumed by the functions, reducing infrastructure costs and overhead. However, we need also to consider the potential drawbacks of Serverless implementation. Consider performance issues, since serverless functions can take longer to start up when they are not frequently used (cold start), leading to slower response times. Also, Serverless in e-commerce can introduce new security risks if not properly configured and monitored.</p><p><strong>Comparison of Architecture styles based on scalability and load needs.</strong></p><p>To keep this simple and easy to remember I have created a tab.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/912/1*WvgYbwyENiXn6FE5gdu9fg.jpeg" /></figure><p><strong>Project — Design of a Scalable web application</strong></p><p>For this task I have to propose a web application, create a traditional server scaling version of it, and also a serverless version. The purpose of this task will be to showcase some practical applications of load balancing, API gateways, and different scaling use cases in practical scenarios.</p><p>Reflecting on what to do, I thought I may do a project of a social network app. So, thinking about what could be something cool for a social networking app, I thought about a platform for connecting people with similar hobbies. Let’s call this application “HobbyHub”. The application will have basic functionalities which are: user management, search, chat and notifications.</p><p>Here below is the traditional server scaling version for this application:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*diJu51IwEHb2nQZUbiwa8w.png" /></figure><p>The incoming requests from the users will pass through the gateway and will be then efficiently distributed by the load balancer across the EC2 instances (which scale vertically automatically). These EC2 will be the front end of the app, split in 2 autoscaling group with an ELB distributing the incoming application traffic to provide performance, availability, and scalability. The reason of going with autoscaling is to get horizontal scaling, where the number of instances is adjusted to meet changing demand. Whereas going just with vertical scaling (which involves increasing or decreasing the hardware resources of the VM) although simpler, it would result in a limit based on the maximum capacity of the server.</p><p>The functions (user management, search, chat, notifications), hence the logic part, will be also on EC2 servers and split in 2 autoscaling groups with an ELB in the middle.</p><p>Regarding the data, there will be a database cluster, supporting vertical scaling, and providing high availability.</p><p>At this point, let’s see how the Serverless version of the app looks:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LcNzb4WpWEfR3zxAD68X1A.png" /></figure><p>In this style we can imagine the frontend to be hosted also on EC2 clusters.</p><p>In this design the API Gateway plays a crucial role because:</p><p>- it effectively protects functions from being exposed directly to the internet;</p><p>- It enforces security policies;</p><p>- it provides authentication and authorization for the users;</p><p>- it manages the inbound traffic by routing the requests to the appropriate Lambda service within the social network app.</p><p>Aside the API Gateway, each application feature is handled by a Lambda function that runs code in response to events and automatically manages the compute resources. Regarding the data layer, here also there is a database cluster same as before as it seems a good solution for this type of app.</p><p>Let’s wrap this up by comparing these two designs. In general, in both cases the users’ requests to the application are processed by the logic part which then involves data retrieval from the database and sending the response back to complete the flow. However, while in the traditional server scaling the servers are always up and running, in the Serverless version the function will be activated only when it’s needed. So, for the traditional version we can have higher resources cost and higher performance while in the Serverless the application will be much more efficient as the functions will be ‘asleep’ when there are no requests, but they may provide a bit of latency when it’s time to ‘wake up’. The scaling also will be simpler in Serverless since happens almost instantly and it is managed by the cloud provider and potentially has no restrictions, whereas a traditional scaling approach involves either provisioning new VMs or increasing hardware resources. This process is slower and more complex.</p><p>All right, I promised to make this simple, so here is a tab comparing and contrasting the aspects for the two architectural approaches of the application.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cxxxK6C-DzYvUYQP9D2q8Q.jpeg" /></figure><p><strong>Project — Design of a Video Sharing Platform</strong></p><p>For this task I have to design a simple video-sharing platform that allows users to upload videos, store video metadata, process videos into multiple resolutions, analyze content for appropriateness, and serve videos to users with minimal latency. The application will feature upload, search, and view functions and should provide also an adult content filter. The objective of the task will be to showcase the interaction between different components like storage, databases, processing services, and the CDN.</p><p>Let’s call this (fantasy) video sharing application “Vidrio”. This came up as a combination of “video” and “rio” (that in Spanish means “river”), and seems a cool name that conveys the idea of stream content flow.</p><p>See the Vidrio project design below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*irSq15IRV0_TWAMD6mDAbA.png" /></figure><p>The landing page will be hosted on a S3 bucket for optimization reasons. Amazon S3 is a popular choice for hosting static websites and landing pages due to its scalability, durability, and cost effectiveness. It is also perfect for media files for the same reasons. The CDN will be a crucial component, ensuring fast, reliable, and scalable video streaming. In terms of security, a CDN provides robust security features, such as encryption and DDoS protection, to safeguard video content and prevent unauthorized access. For searching the metadata an AWS Lambda will be used due to advantage in scalability and cost.</p><p>Here is the action flow when a user uploads a video:</p><p>- the video uploaded will be stored on an S3. As part of the upload process there is the metadata extraction. Metadata is descriptive information associated with the video like title, description, tags, keywords, and other text-based information that describes the video’s subject matter. One can think that storing metadata on a relational database would be ok initially, but imaging that Vidrio could have millions of users uploading at the same time (much like YouTube), a non-relational DB, such as DynamoDB, makes more sense. The reason is because a relational database would struggle to scale horizontally;</p><p>- when the video is uploaded it will be processed to be converted in multiple resolutions to accommodate various types of devices and internet bandwidth speeds. This function could be either a microservice or a serverless function;</p><p>- almost at the same time of the video processing we have the analysis. This will be another service (so can be another function/server) that will leverage machine learning to scan each frame of the video to make sure the content is in line with the platform standards;</p><p>- after that the video will be stored in another S3 bucket and organized based on different resolutions. This process of sorting out can be carried out by a function that will be triggered based on events.</p><p>Let’s see now the flow for watching a video:</p><p>- the user will reach the landing page hosted in S3;</p><p>- the search will reach the CND. At this point, if the content the user was searching for is cached in the CDN, the search is over and the user will watch the content;</p><p>- if the search query is not in the CDN, the request will be sent to the API Gateway which will process it and forward it to a Lambda function, that will get triggered to search for matching metadata based on the search query, and will send that back to the user who will see the list of matching videos;</p><p>- in case the video is cached in the CDN the video will load immediately, otherwise the CDN will fetch the video from the S3 object storage, and will be cached. The next time another user from the same geographical zone will request the same content, will get the content directly from the CDN.</p><p>That is it for this round. Until the next one!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd19c0fbcb67" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My journey in Cloud Engineering: documenting the second module [Git & Cloud Architecture / Intro to…]]></title>
            <link>https://medium.com/@mario.camarca/my-journey-in-cloud-engineering-documenting-the-second-module-git-cloud-architecture-intro-to-f7b9321f1502?source=rss-201c9e8a14a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/f7b9321f1502</guid>
            <dc:creator><![CDATA[Mario Camarca]]></dc:creator>
            <pubDate>Sun, 10 Nov 2024 10:38:26 GMT</pubDate>
            <atom:updated>2024-11-28T20:52:28.647Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>My journey in Cloud Engineering: documenting the second module [Git &amp; Cloud Architecture / Intro to Cloud Architecture].</strong></h3><p>Following up my <a href="https://medium.com/@mario.camarca/my-journey-in-cloud-engineering-documenting-the-first-module-cloud-fundamentals-58116a20d468">first article</a>, in this section we will talk about Version Control systems, explore Git commands, introduce Cloud Architecture concepts, go over load balancing in more detail, and talk about high availability and fault tolerance. Finally, I will show my project of a simple web application architecture.</p><p>A little “disclaimer” before proceeding: as already stated before, I want to remark that in these articles I will not report any part of the study material (which off course is protected by copyright). What you can find, instead, is my own personal summary, which may contain additions and further explanations and/or modifications as I see fit.</p><p><strong>Version Control.</strong></p><p>What is Version control? It is a software engineering practice that enables the tracking, organization, and management of different versions of computer files, source code text files, and other types of files. It maintains a record of every change, and aids in managing changes throughout the software development lifecycle.</p><p>Imagine a big corporate that has to manage a software project, with people working from different places in the world, at different time hours. Version controls allows to collaborate seamlessly, and without worrying that people overwrite each other’s work. It can also save the company from catastrophic errors by allowing to revert back to previous versions. Plus, it creates a clear history of who did what, when, and why. To recap, we can summarize 3 main benefits for using version control:</p><ol><li>Collaboration: it enables multiple developers to work together, track changes, and resolve conflicts.</li><li>Integrity: It ensures that changes are properly recorded and can be reverted if needed.</li><li>Efficiency: it streamlines workflows, reduces errors, and improves communication among team members.</li></ol><p>Before going ahead and explore the main types of version control, we need to stop a moment to check and understand 3 key concepts related to a version control system:</p><p>1. Repository: a database where data, files, or software packages are stored and managed. A repository is used for tracking changes and revisions to the stored items, allowing for collaboration and rollbacks;</p><p>2. Commit: it is a way to capture and store the changes made to a project. It represents a snapshot of the project’s changes at a specific point in time;</p><p>3. Branch: it is a duplication of an object, such as a source code file or a directory tree, allowing for separate and parallel modifications. This enables multiple lines of development to coexist within a software project. By using branches, version control systems (like Git) facilitate collaboration, experimentation, and iterative development, making it easier to manage complex software projects.</p><p>There are 3 types of Version Control systems:</p><p>1. Local: it has database on a local computer, where every file change is stored on the hard drive. This type of system is isolated from external servers or cloud-based repositories.</p><p>2. Centralized: in a CVCS (Centralized Version Control System) the source code is stored in a central server that acts as the main repository. The developers check out the code they need to work on, and make changes directly to that codebase. An example of CVCS is Apache Subversion.</p><p>3. Distributed: unlike the centralized, on a DVCS (Distributed Version Control System) there is not a single central repository. Each developer has a local copy of the repository, and can commit changes and work offline without relying on a central server.</p><p><strong>The most popular and widely used DVCS: Git.</strong></p><p>Git is a powerful, distributed, and decentralized version control system that has become the <em>de facto</em> standard for collaborative software development, offering speed, data integrity, and flexibility. Let’s examine some important git commands.</p><p>Key commands:</p><pre>git init: it creates a new local repository</pre><pre>git clone + url: it clones a repository from an existing url</pre><pre>git status: to check status of changes</pre><pre>git add [file]: used to stage a file for commit</pre><pre>git commit -m “commit message”: used to commit staged changes with a commit message</pre><pre>git push: it is used to upload local commits to a remote repository, such as GitHub or GitLab.</pre><pre>git branch: to list, create or delete branches</pre><pre>git branch [branch-name]: to create a new branch</pre><pre>git checkout [branch-name]: to switch to a different branch</pre><pre>git merge [branch]: to merge a specific branch with the current one</pre><p>Commands to view any git changes:</p><pre>git log: to view commit history</pre><pre>git diff: show changes between commits, branches, etc…</pre><p>Commands to undo changes:</p><pre>git revert [commit]: create a new commit that undoes changes from a specific commit</pre><pre>git reset: reset current HEAD to a specified state</pre><pre>git push [remote-name] [branch-name]: to push a local branch to a remote repository</pre><pre>git pull [remote-name] [branch-name] — to fetch and merge changes from a remote branch</pre><p>Hereunder a screenshot showing git commands interacting with GitHub (a web-based platform that hosts Git repositories) via SSH.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/706/1*RXP-oPQaqcUdofSqzIRMsA.jpeg" /></figure><p>Notice at the beginning after performing “git status” there is nothing to commit. After that I manually added a text line to the README.md file and issued the command again. Git picked up the changes and provided further instructions on what to do next. I went forward with staging the content of the directory (there was only the README.md file) for the change, and then commit. At that point I issued again the “git status” command, hence Git informed that the local branch is ahead of one commit. By issuing the command “git push” the GitHub repository was updated with the changes that were made. At that point, “git status” showed that the branch was up to date, and that there was nothing to commit.</p><p><strong>Introduction to Cloud Architecture.</strong></p><p>Cloud architecture refers to how various cloud technology components combines to build a “Cloud”.</p><p>The main components are the following:</p><p>Frontend: This is essentially what is presented to the user, like for example the UI elements, the visual representation of information, such as text, images, graphics, videos etc… It provides a client-side interface and application for users to interact with. This part can be referred also as “web tier” or “presentation tier”.</p><p>Backend: refers to the server-side part of a web application or program that is not directly accessible to users. It is responsible for processing data, handling requests, and managing interactions between the application and its underlying infrastructure, such as databases, APIs, and operating systems. It enables to manage resources required for cloud computing services. The backend comprehends what is referred also as “logic tier” (e.g. the business logic that processes user inputs, queries databases), and as “data tier” (e.g. a database responsible for storing and managing data).</p><p>Cloud-based Delivery Model: this part is related to what type of cloud service is used, for example SaaS (software as a service), as explained in the previous article. It enables the delivery of cloud services and resources to users.</p><p>Network: this, off course, serves to connect the aforementioned elements as it facilitates communication between cloud components.</p><p>When we are designing our cloud architecture, we want to stick to the following best practices:</p><ul><li>Scalability: this refers to the ability of a cloud infrastructure to increase or decrease IT resources (such as computing power, storage, or bandwidth) as needed to meet changing workload requirements. There are 2 main types of scalability:</li></ul><ol><li>Horizontal (Scale-Out), which involves adding more instances or nodes to a cloud environment to increase processing power or capacity;</li></ol><p>2. Vertical (Scale-Up), which refers to increasing the resources (e.g., CPU, RAM, or storage) of an existing instance or node.</p><p>To recap, horizontal scalability involves for example adding more VMs, while vertical involves increasing processing power of a VM.</p><ul><li>Reliability: this is about ensuring that cloud-based systems and applications are designed to withstand failures, maintain uptime, and recover quickly from disruptions. This concept is related to “high availability” and “load balancing”, which we will examine later on.</li><li>Security: we want to protected our data from unauthorized access/changes, and make sure that the data and the cloud resources in general are always available to legit users. Essentially, we want that confidentiality, integrity and availability principles are protected, and also make sure our organization is compliant with industry regulations.</li><li>Performance: we want to configure the cloud resources to meet performance requirements in order to have optimal service delivery.</li><li>Cost: we want, however, that our implementation is cost effective.</li></ul><p><strong>Load balancing, high availability, fault tolerance.</strong></p><p>A load balancer is a technology designed to distribute the workload between different servers or applications, ensuring efficient and reliable processing of network traffic. In this way, load balancers ensure no single server is overwhelmed, helping to meet users’ demand.</p><p>How this works in practice? Say for example a user connects to Netflix. The client (the user’s laptop, tablet or other device) will send the request to the VIP, which is the publicly accessible address. A VIP is a logical IP address that is mapped to the load balancer, allowing clients to connect to it as if it were the IP address of the actual server hosting the application. The load balancer then receives the request and determines which server from the pool to forward it to. This is based on factors such as server availability, response time, and workload. The selected server processes the request and sends the response back to the load balancer. At that point, the load balancer returns the response to the client, completing the request cycle.</p><p>Load balancing provides benefits in terms of:</p><p>- Scalability: as they add or remove servers as needed to handle changing workloads;</p><p>- Availability: by distributing traffic across multiple servers, load balancers reduce the risk of a single point of failure and improve overall system availability;</p><p>- Performance: load balancers optimize response times and reduce latency by directing requests to the most suitable servers. They also can offload SSL/TLS encryption and decryption processing to dedicated hardware or software, freeing up resources on the backend servers.</p><p><strong>High Availability and Fault Tolerance.</strong></p><p>What is high availability? It is a critical characteristic of modern systems as it ensures the uptime of cloud services. High availability as a feature is commonly seen in the case of cloud service providers, who typically assign a service level agreement (SLA) score to the availability of their cloud systems. For instance, blob storage systems such as Azure, Google Cloud, and AWS S3, all feature an availability SLA of 99.99% per year, which roughly translate to 53 minutes downtime! How high availability is achieved?</p><p>High availability implements several techniques such as:</p><ul><li><strong>Replication and backup</strong>: this is about multiple copies of data stored in different location and regular backups performed;</li><li><strong>Clustering</strong>: in case a server in a cluster fails, the service is uninterrupted as the others (in the cluster) continue to operate;</li><li><strong>Redundant hardware and software</strong>: this involves implementing hardware redundancy such as for example RAID configurations in storage systems, and redundancy in software, for example multiple instances of an application or service running simultaneously to ensure uninterrupted operation.</li></ul><p>If high availability is the ability of a system to handle excessive requests with minimal downtime, what is then fault tolerance? Fault tolerance refers to the ability of a system to function properly and maintain its intended operation even in the presence of failures or errors within its components. Fault tolerance also relies on hardware and software, and also specifically targets disruptions caused by a single point of failure. However, it goes beyond high availability: fault tolerance is about not just making sure the best possible uptime, but also to prevent any downtime.</p><p>For example, a second cluster of servers, a second power generator for a datacenter, an entire data center mirroring the main one located in a different geographical zone etc… these are fault tolerance components.</p><p>To recap, high availability and fault tolerance are two related but distinct concepts in ensuring system resilience. While both aim to minimize downtime and data loss, they differ in their approaches and goals as high availability aims to high uptime and minimal downtime, while fault tolerance aims to zero downtime and zero data loss.</p><p><strong>Design of a basic architecture.</strong></p><p>Enough boring text, it is time to design something! For this task, I have to develop a conceptual design for a cloud infrastructure supporting a web application.</p><p>To do that, I could came up with something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a9CB91o5Wc_fW3WmhH3RiA.png" /></figure><p>Yes, I know, I know… Sure this works, and can be the ideal choice for some temporary/small project. However, there are multiple issues with the above design. Here, all the application tiers are encapsulated in a single Amazon EC2 instance (it’s essentially a VM) that can be accessed directly from the internet. This brings multiple problems as we don’t want all components exposed to the internet for obvious security reasons. Moreover, while vertical scalability could mitigate availability issues that can occur during traffic peaks (multiple users accessing simultaneously), this design is still characterized by a single point of failure. If for some reason the EC2 instance goes down (like for example due to hardware failure, power outage, DDoS attack, etc…) the lack of horizontal scalability and failover will cause the interruption of the service, not to mention the data loss and/or potential corruption.</p><p>Let’s see what we can do to make this better.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JhFtQGBArtA8LoHP8mnshA.png" /></figure><p>In the sketch above the 3 tiers are not incapsulated in one VM. We have the front-end exposed, the back-end is on its own internal subnets, while a load balancer manages the incoming traffic.</p><p>There are autoscaling groups for horizontal scaling to add or remove EC2 instances in order to match changing workload demands. These groups are split in 2 availability zones to prevent that a local disaster may cause service interruption. If for some reason one availability zone becomes unavailable, Amazon EC2 Auto Scaling can launch instances in the other to provide better availability and also fault tolerance.</p><p>As far as concerning the data tier, itr is vertically scalable. This involves increasing the power and resources of a single instance (such as CPU, memory, storage) to withstand the necessity to increase input and output demand. I need to mention, though, that Amazon RDS (as it appears on the sketch) does not support native automatic vertical scaling, hence to avoid doing it manually, in a real scenario, it should be implemented a custom solution using AWS services to allow the automation.</p><p>That’s all for this module, until the next one!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f7b9321f1502" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My journey in Cloud Engineering: documenting the first module [Cloud Fundamentals].]]></title>
            <link>https://medium.com/@mario.camarca/my-journey-in-cloud-engineering-documenting-the-first-module-cloud-fundamentals-58116a20d468?source=rss-201c9e8a14a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/58116a20d468</guid>
            <dc:creator><![CDATA[Mario Camarca]]></dc:creator>
            <pubDate>Fri, 25 Oct 2024 17:29:42 GMT</pubDate>
            <atom:updated>2024-11-04T17:06:49.882Z</atom:updated>
            <content:encoded><![CDATA[<p>I am starting this blog documenting my journey in cloud engineering to help myself in this new endeavor, and also (hopefully) help anybody else on the same path, or simply passionate about tech, that will stumble across this article. To achieve this, I won’t just summarize the concepts and topics in the study material, and show part of lab/exercises, but I will also try to give some examples along with my personal comments and explanation.</p><p>Before proceeding a little… I would say “disclaimer”. I want to remark that in these articles I will not report any part of the study material (which off course is protected by copyright). What you can find, instead, is my own personal summary, which contains additions and sometimes explanations as I see fit.</p><p>Besides, name’s Mario, PC enthusiast, and cybersecurity professional with experience ranging from malware analysis and digital forensics to SOC and incident response tasks.</p><p>Cloud Computing is a very hot topic indeed, and I believe it is a very important one for a security practitioner.</p><p>Ok then, let’s start talking about the material of the study module “Cloud Fundamentals”, exploring cloud computing, its deployment models, the technologies involved, and other topics related to it.</p><p><strong>What is Cloud computing?</strong></p><p>Cloud computing is the on-demand delivery of computing services (servers, storage, databases, networking, software, analytics, AI…) over the internet. This model enables private and business users to access and utilize IT resources without the need for upfront capital expenditures or management of physical infrastructure.</p><p>If this concept sounds like some boring abstract stuff already heard/seen somewhere else, then just think in reality there is no “Cloud”. It is just computer hardware somewhere else.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/626/1*czD3-oOHgNUPSA5F1EsTbQ.png" /></figure><p>The largest cloud providers are Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP).</p><p>Key characteristics.</p><p>On-demand access: Resources are provisioned and de-provisioned as needed, without the need for manual intervention.</p><p>Pay-as-you-go pricing: Users only pay for the resources consumed, reducing capital expenditures and operational costs.</p><p>Scalability and Elasticity: The first means resources can be quickly scaled up or down to match changing business needs, and the second means that this involves automatic adaptation.</p><p>Multi-tenancy: Resources are shared among multiple users, with each user’s data and applications isolated from others.</p><p>Main deployment models.</p><p>Public Cloud: A public cloud is a multi-tenant environment where resources are owned and managed by the provider, accessible over the internet. It is an easy and scalable cost-effective solution, but offers less security and control. It is suitable for businesses with fluctuating demands and low-security concerns.</p><p>Private Cloud: A private cloud is a single-tenant environment where resources are owned and managed by the organization itself, either on-premises or through a third-party provider. This model offers higher security and control, it is customizable and tailored to specific needs, and offers scalability and flexibility. It is suitable for organizations requiring high security, compliance, or customization.</p><p>Hybrid Cloud: A hybrid cloud combines public and private cloud environments, allowing data and applications to be shared between them. This model balances security and scalability, enables workload migration between public and private clouds. It is suitable for organizations with diverse workloads and requirements.</p><p>Types of cloud service models.</p><p>Infrastructure as a Service (IaaS): Provides virtualized computing resources, such as servers, storage, and networking. In this case, we (as a business or just users) would be responsible of everything else, such for example choosing an OS, perform its installation, configuration etc… Some examples of IaaS are AWS, Microsoft Azure, and Google Compute Engine.</p><p>Platform as a Service (PaaS): Offers a complete development and deployment environment for applications, including tools, libraries, and infrastructure. Here we have a pre-configured environment already, we have to add all programs and applications. Examples of PaaS are AWS Elastic Beanstalk, Google App Engine, and Adobe Commerce.</p><p>Software as a Service (SaaS): Delivers software applications over the internet, eliminating the need for local installation and maintenance. Hence, we will be responsible only of the data which will be either manually entered or automatically pushed from our other applications to the SaaS application. Examples of SaaS are Gmail, Slack, Dropbox, and Microsoft Office 365.</p><p>What are the advantages of cloud computing?</p><p>We want to use cloud computing to build or outsource all or some part of our IT infrastructure, storage, and applications instead of having everything on premises for multiple practical reasons. In a business perspective these are mostly efficiency/cost related.</p><p>As an Italian, I think the image below (found on the internet) showcase really well the concepts above :)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*xP5akpEe_z1oInIBmQ43bg.png" /></figure><p>How all these things can be achieved? Let’s have an overview at the technologies involved, and essential concepts related to them.</p><p><strong>Virtualization.</strong></p><p>If you ever tried messing around with VMware Workstation or Oracle VirtualBox, maybe to try some fancy Linux distribution, for testing, or any other reason, then you already have an idea what is virtualization. This technology allows multiple operating systems (OSs), to run on a single physical hardware platform. An operating system is the main software program that acts as an intermediary between the user and computer hardware, and manages the computer’s resources, allowing us (users) to interact with the machine.</p><p>Hypervisors are software layers that enable virtualization, allowing multiple virtual machines (VMs) to run on a single physical host. There are two primary types of hypervisors: Type 1 and Type 2.</p><p>Type 1 hypervisors are installed directly on the hardware, while Type 2 hypervisors are installed on top of an existing OS (just like in the situation where we are playing with VMware Workstation or Oracle VirtualBox).</p><p>Type 1 hypervisors are generally faster and more efficient due to direct hardware access, they can implement custom resource allocation strategies, and therefore are ideal for large-scale, resource-intensive workloads such as in enterprise environments. Type 2 hypervisors, on the other hand, are better suited for desktop and development environments.</p><p>Examples of type 1 hypervisors are VMware ESXi, Microsoft Hyper-V, KVM. For type 2 hypervisors, examples are VirtualBox VMware Workstation, off course.</p><p>After we have set up our virtual machine, we need to chose and install an operating system (just like we would do on a fresh PC build).</p><p>Virtualization not only allows to create virtual servers or workstation but it applies also to networking and storage. It is possible to create a logical network overlay on top of a physical network infrastructure. This allows multiple virtual networks (VNs) to coexist and operate independently.</p><p>The same concept is valid for storage virtualization. This technology abstracts physical storage devices and presents them as a single, centralized pool of storage capacity, allowing organizations to manage multiple storage arrays and devices from different type (mechanical hard drives as well as SSDs) and vendors as a single, unified storage system.</p><p>Virtualization provides lots of benefits in terms of:</p><p>Scalability: easy scaling of resources;</p><p>Cost effectiveness : no need to invest in a data center;</p><p>Flexibility: VMs can be easily created, cloned, and migrated;</p><p>Security : VMs are isolated from each other and the physical host, reducing the attack surface;</p><p>Disaster recovery : virtualization enables quick recovery from failures or disasters by simply restarting VMs on a different host.</p><p>However, it brings also cons:</p><p>Complexity: Implementing and managing virtualization requires specialized skills and knowledge;</p><p>Performance overhead: Virtualization introduces additional layers of abstraction, which can impact performance and require optimization;</p><p>Resource contention: Multiple VMs competing for resources can lead to performance degradation and contention;</p><p>Security risks: While virtualization provides isolation, it also increases the attack surface and requires careful configuration and monitoring to mitigate risks.</p><p><strong>Basic networking concepts and protocols.</strong></p><p>Networking is the process of connecting two or more computing devices (such as desktop computers, mobile devices, routers, or applications), to enable the transmission and exchange of information and resources. In essence, networking allows us to create a system of interconnected devices that can share information, resources, and services, enabling communication, collaboration, and data exchange between them. To connect multiple networks we would use routers, while to connect devices within a single network we would use switches.</p><p>In a network, devices communicate with each other using standardized protocols. See below a breakdown of some key protocols.</p><p>TCP/IP (Transmission Control Protocol/Internet Protocol).</p><p>TCP/IP is an essential suite of communication protocols used to interconnect devices on the internet. It is used to manage and route packets, and ensures reliable, error-checked, and sequential data transfer between devices.</p><p>TCP (Transmission Control Protocol) and UDP (User Datagram Protocol).</p><p>As mentioned above, TCP ensures reliable transmission by segmenting data, tracking segments with sequence numbers, and using acknowledgments and flow control to manage data transfer. Its error detection mechanisms and connection establishment process make it a fundamental protocol for internet communication. TCP is best used for direct communication in which a reliable connection is needed, such as web browsing, and email.</p><p>On the other hand, UDP (User Datagram Protocol) is a connectionless protocol. UDP provides best-effort delivery without guaranteeing packet delivery or order, making it suitable for applications that require fast transmission but can tolerate some packet loss or corruption. An example of UDP use case is for time-sensitive transmissions such as video playback.</p><p>IP (Internet Protocol).</p><p>IP assigns unique addresses, known as IP addresses, to devices on a network and serves to route data between nodes. An IP address can be in IPv4 format which is 32-bit addresses (example: 192.168.1.5), allowing 4 billion possible addresses, or in IPv6 format which then it is a 128-bit address (example: fd07:a47c:3742:823e:3b02:76:982b:463), with 340 undecillion possible addresses.</p><p>An IP address can be private or public. The first type is not routable on the Internet, and it is used for communication within a local network. The latter is routable on the Internet and can be accessed directly.</p><p>A crucial component in IP addressing is the subnet mask, as it allows dividing a network into smaller, more manageable subnetworks and also indicates to what network a given IP address belongs to.</p><p>DHCP (Dynamic Host Configuration Protocol).</p><p>An IP address can be either static, for example it was manually assigned to a device (permanently) by the network administrator for some purpose, or can be dynamic, which is when it is temporary and so can change frequently. In this regard, DHCP is used to assign IP addresses (and other network settings) to devices automatically from a pool of available IPs. An example of DHCP usage is when we connect to internet, as our computer will get an IP assigned from the internet provider DHCP server. Another example is when a smartphone connects to the home Wi-Fi network, as the home router is acting as a DHCP server.</p><p>ICMP (Internet Control Message Protocol).</p><p>The Internet Control Message Protocol (ICMP) is a network layer protocol used by network devices to diagnose network communication issues. Ping is a diagnostic tool that uses ICMP, it works by sending Internet Control Message Protocol (ICMP) Echo Request packets to a specified destination and then waiting for a reply to determine: Packet loss (percentage of packets not received), Round-trip time (RTT) in milliseconds, Minimum, maximum, and average RTT.</p><p>We can use ping to test network connectivity to a destination host or device. See below a screenshot of ping usage as performed in a testing assignment of this module.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/534/1*NgEjX6J-ofXVF_Yf1oFwow.png" /></figure><p>Tracert/Traceroute.</p><p>Tracert (the command in Linux is traceroute) is another diagnostic command-line utility. It displays the path and transit times of packets across an IP network to help us troubleshoot connectivity issues and identify bottlenecks in network communication.</p><p>See in the screenshot below how the tool shows all the hops the packets traveled through to reach the destination. Notice the stars at point 4, suggesting there was no response and the request was timed out.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/1*746-8dMMpLCez9fi_2Fodg.png" /></figure><p>DNS (Domain Name System).</p><p>DNS is a hierarchical and distributed naming system that translates human-readable domain names into IP addresses. We can think of it as the phone book of the internet.</p><p>When we type “google.com” in the browser bar a series of actions are triggered: in case the record is not in the machine cache, a request will be sent to the recursive resolver server (it is called like that cause it will perform multiple requests). This will send a request to the root server which will respond with the TLD (top level domain, such as “.com”, “.org”, “.io”…) server address. The resolver then will make the request there and will get the address of the authoritative server (in our example the authoritative server will be google authoritative name server). The resolver will reach out the authoritative and fetch the IP address. At this point the address will be provided to client that finally can connect to google.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/613/1*B7jXsAxxoNgZsmq71bITbA.png" /></figure><p>HTTP/HTTPS protocols.</p><p>HTTP (Hypertext Transfer Protocol) and HTTPS (Hypertext Transfer Protocol Secure) are protocols used for transferring data over the internet, using TCP/IP (Transmission Control Protocol/Internet Protocol) as underlying transport protocol.</p><p>In the assignment I am using the “curl” command to make GET requests to retrieve HTML information from a website. With “curl -I” we can view and analyze HTTP response headers, as shown in the screenshot below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/461/1*ibZvv6DjJHHt3hbNO7N_hg.png" /></figure><p>As it is visible on the screenshot, the HTTP status code was 200 (OK), which means the request was successful. There are many status codes that can indicate the outcome of a client’s request. Here I am going to mention the most common ones that most likely everybody is familiar with, like 404 (Not Found), which occurs when a client (usually a web browser) requests a resource from a server but the server cannot find the requested resource; 301 (Moved Permanently) meaning the requested resource has been relocated to a new URL, which is provided by the location header in the response.</p><p>To have an idea at glace when looking at HTTP codes, as mnemonics, we just need to remember: 2xx = Successful; 3xx = Redirection; 4xx = Client Error; 5xx = Server Error.</p><p>I want to mention also that during this lab exercise I have tried to curl a known e-commerce website as a test, and I got HTTP 301 response. The solution to make curl follow the redirection was to add the option -L.</p><p>Before ending documenting this section dedicated to basic networking concepts and protocols, need to mention also SMTP (Simple Mail Transfer Protocol), and FTP (File Transfer Protocol). SMTP is a protocol for exchanging email messages between servers while FTP is used for transferring files between systems over a network. Not to stretch this too much, but it is worth noting that these protocols don’t have security built in, therefore it is advised to use the secure version of them, especially in untrusted networks (such as the internet!).</p><p><strong>Key components and technologies in a cloud network.</strong></p><p>Virtual networks.</p><p>Virtual networks are software defined network that emulate physical network functionality, they connect virtual machine and other cloud resources. A practical example can be a business using SDN to create three different environments: one for developing, one for testing, and one for production. Even though not related to this specific topic, mentioning “developing” and “testing” reminded me another concept covered in the present module which is SDLC (Software Development Life Cycle), a process used to design, develop, test, and maintain high-quality software systems. It consists of various stages, which may vary depending on the methodology or model use. There are usually six stages in the cycle: planning, requirement analysis, design, development and testing, deployment, maintenance. The most popular models are “Waterfall”, that has a linear sequence of stages, and “Agile”, that has and iterative and incremental approach instead.</p><p>It is not by chance that this was mentioned already in the first module. A cloud engineer can add significant value in supporting the SDLC process.</p><p>Routers and load balancers.</p><p>Cloud environment routers manage traffic between the cloud resources and the outside (internet), they provide connectivity, packet filtering, path determination in order to optimize the connection; much like as they work in non cloud environments.</p><p>Load balancers are used to distribute capacity during peak traffic times, and to increase reliability of applications.</p><p>Imagine a load balancer as a physical or virtual appliance (hardware-based, or software-based) that sits between the user and the server group, and acts as an invisible facilitator ensuring that all resource servers are used equally and provide availability for all users.</p><p>For example, let’s think about network traffic surging during Christmas time toward, say Amazon. Load balancers will handle the requests making sure the website will be available for all the users.</p><p>Internet gateways and VPNs.</p><p>An internet gateway allows the cloud network to exit to the internet and access to the cloud resources from outside. Essentially acts as a link for the cloud environment and the internet.</p><p>VPNs are used for secure remote access to the cloud resources. Having remote users using a VPN to access the cloud resources, from the outside, will ensure the connection is secure as it is encrypted. This will protect the traffic against potential eavesdrop attacks.</p><p>Firewall and Encryption.</p><p>A firewall is a network security device that monitors and controls incoming and outgoing network traffic based on predetermined security rules.</p><p>Encryption is the process of converting plaintext data (readable information) into ciphertext (unreadable information) to protect it from unauthorized access. With encryption we can protect data in transit (traveling across the network), as well as and data at rest (data stored in a resource).</p><p>Connectivity models in Cloud Networking.</p><p>Direct connect.</p><p>This provides a direct private connection from the company’s on premises infrastructure to the cloud service. The advantages come in terms of stability and performance, with disadvantages in terms of flexibility.</p><p>An example of this is AWS Direct Connect: a dedicated network connection between the organization premises and AWS, providing a private, high-bandwidth link for data transfer.</p><p>Hybrid networking.</p><p>This model involves connecting on premises infrastructure with cloud resources, providing the flexibility of the cloud while maintaining a tighter control of the on premises resources. An example of this could be a company using hybrid networking to connect its on premises database to cloud based applications.</p><p>As I’ve mention databases, need to dedicate a section for them as it is part of this first module as well!</p><p>Databases.</p><p>A database is a collection of organized data stored in a system, structured in a way that allows efficient querying and processing.</p><p>Databases are important parts in web applications and part of the backend. They can store data related to users such as usernames and passwords, and data relating to applications.</p><p>A database is also a crucial component in software development, serving as a centralized repository for storing, organizing, and retrieving data.</p><p>There are many types of databases, the main ones are relational and non-relational databases.</p><p>Relational Database: these databases store data in tables with defined relationships between them, using a structured query language (SQL). Examples include MySQL, PostgreSQL, and Microsoft SQL Server.</p><p>Non-relational database: also known as a NoSQL database, is a type of database that does not use the traditional table-based relational model used in relational databases. Instead, they employ various data models, such as key-value, document, graph, and column-family stores, optimized for specific use cases and data structures.</p><p>Relational databases are suitable for applications with structured data and complex queries. However, they offer less flexibility. Non-relational databases, instead, are ideal for big data, real-time web applications, and content management systems requiring flexible data models.</p><p>It is time to conclude the documentation of this first module. It was really nice to review these concepts and refresh some hands-on command line tools! Until the next one!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=58116a20d468" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>