Zachary Goldstein’s Post

McDonald's food delivery processes over 20,000 orders per second & serves over 64,000,000 customers daily. All the orders are processed with less than 100 milliseconds of latency across 37,000 locations. I spent my weekend reading Neo Kim’s article, and this is what I learned: ↪ Key Components: They use a hexagonal and event-driven architecture that helps McDonald's food delivery platform maintain: - low complexity - high performance - robust scalability, - and handling massive order volumes efficiently. 1. Hexagonal Architecture: - Separates core applications from external services like databases and UI. - Business logic is inside the hexagon, interacting with external services through ports. - Adapters implement these ports to fulfill the required functionality. 2. Event-Driven Architecture: - Uses events for communication, ensuring modularity and loose coupling. - Events are published to a message broker and consumed by relevant services. - A schema registry maintains contracts for events, ensuring consistency and validation. ↪ Order Processing Steps: 1. Choosing Food From the Menu: - Users access a reverse proxy server with all API endpoints. - Menus and hours are stored in an SQL database and served as JSON via REST APIs. 2. Getting a Discount With Loyalty Rewards: - User events are published to a message queue for asynchronous processing. - Serverless functions poll the queue to apply loyalty rewards and discounts. 3. Ordering Food: - A WebSocket connection allows bidirectional communication. - Orders are validated via an event bus and checked for availability with serverless functions. - Redis processes orders for low-latency caching, backed by a relational database. 4. Giving Feedback: - Feedback is collected from surveys and social media, stored in S3, and analyzed with NLP. - The processed data is stored in a data warehouse for reports. ↪ Scalability and Reliability: - Supports 37,000+ locations and 64 million daily customers. - Uses microservices with varying scaling and runtime profiles. - Prioritizes critical customer-facing services; background services can tolerate failures. - Employs smoke testing, circuit breakers, and exponential backoff for resilience. - Extensive automation reduces operational efforts. – P.S: Next time you order a Big Mac, remember how your order is being processed :)

McDonald's food delivery processes over 20,000 orders per second & serves over 64,000,000 customers daily. All the orders are processed with less than 100 milliseconds of latency across 37,000 locations. I spent my weekend reading Neo Kim’s article, and this is what I learned: ↪ Key Components: They use a hexagonal and event-driven architecture that helps McDonald's food delivery platform maintain: - low complexity - high performance - robust scalability, - and handling massive order volumes efficiently. 1. Hexagonal Architecture:    - Separates core applications from external services like databases and UI.    - Business logic is inside the hexagon, interacting with external services through ports.    - Adapters implement these ports to fulfill the required functionality. 2. Event-Driven Architecture:    - Uses events for communication, ensuring modularity and loose coupling.    - Events are published to a message broker and consumed by relevant services.    - A schema registry maintains contracts for events, ensuring consistency and validation. ↪ Order Processing Steps: 1. Choosing Food From the Menu:    - Users access a reverse proxy server with all API endpoints.    - Menus and hours are stored in an SQL database and served as JSON via REST APIs. 2. Getting a Discount With Loyalty Rewards:    - User events are published to a message queue for asynchronous processing.    - Serverless functions poll the queue to apply loyalty rewards and discounts. 3. Ordering Food:    - A WebSocket connection allows bidirectional communication.    - Orders are validated via an event bus and checked for availability with serverless functions.    - Redis processes orders for low-latency caching, backed by a relational database. 4. Giving Feedback:    - Feedback is collected from surveys and social media, stored in S3, and analyzed with NLP.    - The processed data is stored in a data warehouse for reports. ↪ Scalability and Reliability: - Supports 37,000+ locations and 64 million daily customers. - Uses microservices with varying scaling and runtime profiles. - Prioritizes critical customer-facing services; background services can tolerate failures. - Employs smoke testing, circuit breakers, and exponential backoff for resilience. - Extensive automation reduces operational efforts. – P.S: Next time you order a Big Mac, remember how your order is being processed :)

API Types: A Simple Guide with Real-World Examples 🔍 What is an API? Think of an API as a waiter at a restaurant. You (the client) tell the waiter (the API) what you want (data or services), and the waiter brings it to you. That's how APIs help apps talk to each other. Let’s break down the most common API types using daily scenarios: 🌐 REST API (Representational State Transfer) Example: Ordering food from a menu. How it works: You tell the waiter which dish you want (specific data or resource), and they bring it to you. Real-world use: Fetching user profiles from a social media app. 📊 GraphQL Example: Building your own meal with exactly what you want. How it works: You ask for only the ingredients you need, and GraphQL serves it all in one go. Real-world use: Requesting a list of products with specific details like names, prices, and images from an e-commerce platform. 🧾 SOAP (Simple Object Access Protocol) Example: Ordering food using a strict and formal process. How it works: Uses a detailed, standardized method (usually XML) to communicate. Real-world use: Large enterprise systems like banking or payment gateways. ⚡ gRPC (Google Remote Procedure Call) Example: A quick, efficient messaging system in a busy kitchen. How it works: Sends data in an optimized binary format for fast communication. Real-world use: Connecting microservices in applications that need high speed and efficiency. 💬 WebSocket Example: A live conversation between you and a friend, where messages flow instantly. How it works: Enables two-way, real-time communication between client and server. Real-world use: Chat apps, real-time stock market updates, or online gaming. 🌍 Open/Public APIs Example: A restaurant's public menu that anyone can view and order from. How it works: Accessible to anyone on the web. Real-world use: Popular APIs from Google Maps, Twitter, or Stripe for payments. 🔒 Internal/Private APIs Example: A private communication channel within the kitchen staff. How it works: Used internally within a company or organization. Real-world use: APIs for handling internal data, operations, or microservices that aren't exposed to the public. 🗄️ Database APIs Example: A waiter who knows how to manage the restaurant's stock of ingredients. How it works: Used to perform actions like creating, reading, updating, or deleting data in a database. Real-world use: APIs that manage user data or inventory in applications. 🛠️ Which API is Right for You? The choice of API depends on several factors: Data needs: How much and what kind of data do you need? Performance: How quickly does your system need to respond? Security: How sensitive is the data being shared? Scalability: Can the API handle a lot of users or traffic? By understanding these API types, you can pick the best one for your project's specific needs! #API #WebDevelopment #Tech #APIDevelopment #GraphQL #REST #gRPC #WebSockets

Restaurant owners are drowning in too many sales tech tools. They have to navigate 4+ tools just to sell, manage gift cards, and create reports. And these tools don't talk to each other. So, reporting is inaccurate when you manually reconcile all the different data. I hear it all the time from owners: They're frustrated and tired that they can't spend more time being proactive. That's why I've discovered a streamlined solution for data management. A shift from fragmentation to integration. From juggling multiple platforms to a unified system. Here's what I’ve seen work: Centralized Data Platform: --> Use a system that consolidates all your data in one place. --> No more downloading and reconciling from multiple sources. Automated Data Sync: --> Implement software that automatically syncs data across systems. --> Save time and reduce errors with real-time updates. Actionable Insights: --> Leverage integrated data to make informed decisions. --> Focus on marketing, customer retention, and product improvement. This approach has transformed data management for many restaurants. And shifted their focus from administrative tasks to growth strategies.

𝑫𝒊𝒔𝒄𝒐𝒗𝒆𝒓 𝒕𝒉𝒆 𝑭𝒖𝒕𝒖𝒓𝒆 𝒐𝒇 𝑫𝒂𝒕𝒂 𝑺𝒉𝒐𝒑𝒑𝒊𝒏𝒈! Let's compare our technical data products with dishes in a shop like Lieferando. How do these (data) products make their way from creation to you? Let's break down the steps leading to Data Shopping: 💎 𝐃𝐚𝐭𝐚 𝐚𝐬 𝐏𝐫𝐨𝐝𝐮𝐜𝐭𝐬 In the first step, raw data is transformed into valuable data products through 𝐃𝐚𝐭𝐚 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠. Just like a kitchen where raw ingredients are turned into delicious meals, each data point undergoes a meticulous production process. 🛒 𝐓𝐡𝐞 𝐌𝐞𝐧𝐮 and 𝐭𝐡𝐞 𝐌𝐚𝐫𝐤𝐞𝐭𝐩𝐥𝐚𝐜𝐞 Once the data products are ready, they are cataloged and listed on a marketplace – much like an online food ordering platform. Here, users can easily browse through the available data and select the products that meet their needs (e.g. vegetarian, gluten-free etc.). 🚚 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐞𝐝 𝐃𝐚𝐭𝐚 𝐃𝐞𝐥𝐢𝐯𝐞𝐫𝐲 Now comes the innovative part: 𝐃𝐚𝐭𝐚 𝐀𝐜𝐜𝐞𝐬𝐬 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 and 𝐃𝐚𝐭𝐚 𝐕𝐢𝐫𝐭𝐮𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧. These technologies ensure automated data delivery. Instead of relying on manual processes by a Data Engineer, an automated delivery service – similar to Lieferando's couriers – handles the task of delivering data products quickly and efficiently to users. This approach seamlessly and efficiently handles the entire process from data creation to delivery. It's like ordering your data from Lieferando and having it delivered to your door in no time. 💡 Want to learn more about data shopping? Join me and my colleague Georg Frey for our presentation at 𝐓𝐃𝐖𝐈 𝐌ü𝐧𝐜𝐡𝐞𝐧 2024 𝐨𝐧 𝐉𝐮𝐧𝐞 13𝐭𝐡 𝐚𝐭 15:10, where we will present "𝐒𝐭𝐫𝐚𝐭𝐞𝐠𝐢𝐞𝐬 𝐟𝐨𝐫 𝐚 𝐒𝐞𝐚𝐦𝐥𝐞𝐬𝐬 𝐃𝐚𝐭𝐚 𝐒𝐡𝐨𝐩𝐩𝐢𝐧𝐠 𝐄𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞." #datashopping #TDWI #datamanagement #datacatalog #datavirtualization #dataengineering

1. Speedy Delivery vs. Lazy Sunday: The Data Dash In the world of quick commerce, real-time data ingestion is the secret sauce. Imagine a pizza place that tracks every order, adjusting inventory on the fly. Now, compare that to a quaint bakery that checks its sales at the end of the day. Quick commerce needs the speed of a cheetah; traditional businesses can take a leisurely stroll. A solid data pipeline ensures the cheetah never stumbles! 2. Flash Sales vs. Slow Cooked Meals: Timing is Everything Think of a flash sale website that relies on real-time data to grab those impulsive buyers, versus a gourmet restaurant that plans its menu weeks in advance. Quick commerce thrives on immediate insights, while some businesses are perfectly happy with a slow simmer. A well-structured data pipeline makes sure the flash sale doesn’t turn into a fizzle! 3. The Race to the Checkout vs. The Couch Potato For e-commerce giants, every millisecond counts at checkout. Now picture a bookstore that updates its stock once a week—very chill. Quick commerce needs the hustle, while others enjoy the Netflix-and-chill approach. With an efficient data pipeline, the checkout race remains a sprint, not a crawl! 4. The Instant Coffee vs. The Perfect Brew: Data at Your Service Quick commerce is like instant coffee—ready in a flash! Meanwhile, some brands take their time, savoring the brewing process. Real-time data ingestion lets businesses serve customers instantly, while others can afford to take a coffee break. A robust data pipeline ensures that even instant coffee is brewed to perfection! 5. The News Flash vs. The Evening Paper In the fast-paced world of quick commerce, timely insights are like breaking news. Contrast that with a business that waits for the evening paper to learn what’s trending. Some need to stay ahead of the curve, while others are just fine with catching up later. A reliable data pipeline keeps those news flashes coming, ensuring no one’s left behind! In a world of rapid change, choosing the right data ingestion strategy is key to staying relevant. Whether you need speed or a more relaxed approach, a solid data pipeline can help you find your rhythm! #DataIngestion #RealTimeData #QuickCommerce #DataPipelines #BusinessIntelligence #Ecommerce #Analytics #DigitalTransformation #DataDriven #TechTrends #Innovation #CustomerExperience #DataStrategy #SupplyChain #AgileBusiness

It’s been 4 years since Wine Hub was first released, initially driven by wine merchants who felt their needs were unmet, or seriously compromised. Why were we lobbied to create an industry standard business management system? Wine and spirits SMEs don’t generally have the skills to specify and manage a custom project. Skills that in any case aren’t required when you’re part of an industry wide roadmap. What’s the advantage of being sector specific? No design compromises mean: ✅ more efficient operations ✅ workflows that are far more logical than those those adapted from a common-to-all underlying framework ✅ industry specific integrations that work out of the box, ✅ and of course no ongoing development fees. No time wasted trying to explain what you need. ⛔ Are there any disadvantages? Not any more. Being industry specific means doing all the things that really matter to support operational improvements, more sales and paths to growth. Being industry specific also means not trying to build application solutions that don’t need to be industry specific. Think accounting, CRM, eCommerce, point of sale , business analysis tools and logistics. A decade ago integrating all those pieces was tricky, now API integrations are ubiquitous and reliable - and open up a whole world of best in class software that doesn’t lock you in and that also contribute to driving your business forward. What do businesses want from their tech? 👉 Software that doesn’t get in the way, or slow them down, and that frees up team members to do more of what they do best. 👉 Software that helps them to sell more and put smiles back on customers’ faces by delivering great customer service both pre sales and at every point thereafter. Curious about the impact of sector-specific software? Read more here: https://lnkd.in/e5PE_rXM #wine #winetech #winesoftware #wineindustry

Are you struggling to optimize your food and beverage supply chain? 🍽️ Limited visibility, manual processes, and costly production runs can hinder your operational efficiency. But fear not, I've got the solution! Join me as I share exclusive insights on overcoming these challenges and driving organizational efficiencies in the F&B sector. Discover how InterSystems Supply Chain Orchestrator™ empowers real-time intelligent decision-making and streamlines fulfillment processes. Say goodbye to sub-optimal production runs and hello to operational excellence! Read my section of the article here: https://lnkd.in/eYsCqy2C #supplychain #foodandbeverage #dataanalytics FoodChain Magazine

Struggling to optimize your supply chain? Limited visibility, manual processes, and costly production runs can be major obstacles. Don't miss Mark S. Holmes's exclusive insights on driving organizational efficiencies in the F&B sector. Learn how InterSystems Supply Chain Orchestrator™ revolutionizes decision-making and streamlines fulfillment processes in real-time. Dive deep into optimization. Join us now! #supplychainoptimization #foodandbeverageindustry #efficiency

Are you ready to unlock the power of real-time data streaming and event processing? Let's talk about Kafka! Publish-Subscribe Messaging System: Kafka follows a publish-subscribe model where producers publish data to topics, and consumers subscribe to those topics to receive the data. This decouples producers and consumers, allowing for flexible and scalable communication between various components of a system. Real-time Processing: Need to analyze data as it arrives? Kafka's low-latency design enables real-time processing, empowering businesses to make informed decisions at the speed of thought. Imagine you're craving your favorite pizza 🍽️ from the local pizzeria. Here's how Kafka adds a dash of magic to your pizza experience: 🛵 Order Placement: You place your order through the pizzeria's app. Each order is like a delicious slice of data, ready to be processed. 🍕 Kitchen Coordination: Kafka acts as the kitchen manager, ensuring that every order is received promptly and assigned to the right chef. Just like topics in Kafka, each order gets its own slice of attention. ⏱️ Real-time Baking: While you eagerly await your pizza, Kafka orchestrates the baking process in real-time. From dough to toppings, each step is synchronized to perfection. 📱 Delivery Updates: As your pizza gets baked, Kafka keeps you in the loop with real-time delivery updates. Just like tracking a package, you can follow your pizza's journey from oven to doorstep. 📊 Feedback Loop: After indulging in your pizza delight, you share feedback through the app. Kafka captures your comments, allowing the pizzeria to improve their service based on real-time insights. With Kafka, every pizza order becomes a seamless, enjoyable experience, from click to crust! 🍕✨ #ApacheKafka #RealTimeDelivery #PizzaMagic #DataStreaming Hungry for more insights on how Kafka can transform your business? Let's chat! 🌟📈