June 2022 Summaries

DataStax has announced the general availability of Astra Streaming, a high-scale messaging and event streaming service with built-in support for Apache Kafka, RabbitMQ, and JMS. This service, which is built on open-source Apache Pulsar™, combined with DataStax Astra DB, their managed database service built on Apache Cassandra®, creates an open stack that unifies all aspects of real-time data: both operational data-at-rest and streaming data-in-motion. This cloud-native stack allows organizations to mobilize all enterprise data for real-time applications, build smarter applications faster, and scale without limits on any cloud. The technology enables businesses to deliver in-the-moment experiences to customers, create instant intelligence, or fuel machine learning, which is critical for staying relevant in today's digital landscape.

Real-time data is becoming increasingly important as companies strive to ingest, process, and act on data in shorter iterations to improve customer experience and drive efficiencies. Event streaming plays a crucial role in real-time use cases such as fraud detection and intrusion detection. However, legacy messaging platforms often fail to capture these events effectively. Astra Streaming is a cloud-native event streaming service built on Apache Pulsar that provides a drop-in replacement for Kafka, RabbitMQ, and JMS, allowing enterprises to quickly write code for in-stream processing of their event data. With its integration with Astra DB's change data capture capabilities, it offers the industry's first solution that unifies data-at-rest and data-in-motion across clouds.

Cloud native architecture is a modern approach to building applications that leverages the capabilities offered by cloud infrastructure providers and specialized cloud-based software services. It involves developing and deploying independent, self-contained microservices, which can be sourced from different providers or developed in-house. The key principles of cloud native architecture include consistent and reusable APIs, stateless applications for easier scaling, domain-driven design for clear communication among teams, and automation to manage the complexity of cloud resources. Benefits of this approach include on-demand scalability, streamlined updates, and more time for refining applications after launch.

Starlight for JMS is a powerful API built on Apache Pulsar that helps developers create reliable, easy-to-scale, and fast applications. It implements both JMS 1.1 and JMS 2.0 specifications and can be used by simply including its Java library dependency in your application and replacing the legacy JMS dependency. Starlight for JMS stores messages in an Apache Pulsar cluster and offers several benefits to JMS applications, such as robust authentication mechanisms like JSON Web Token (JWT) based authentication. The connection properties fall into several categories: related to establishing a connection to the Pulsar cluster, related with Starlight for JMS API behavior itself, and related with fine-tuning Pulsar producer and consumer behaviors. Starlight for JMS also supports JNDI lookup and implements JMS message payload types by mapping them to Pulsar BYTES message type. It provides support for JMS message producers, consumers, queue browsers, and the Request-Response pattern. The code changes required to upgrade existing JMS applications to Starlight are minimal due to its compatibility with JMS spec 2.0 and backward compatibility with JMS spec 1.1.

Serverless databases-as-a-service (DBaaS) offer numerous benefits over traditional on-premises servers, including eliminating capacity headaches, improving service availability and performance. By matching serverless functions with a serverless NoSQL database, developers can power up app performance while driving down costs. A managed service like serverless DBaaS frees organizations to focus on speed, performance, and simplified operations, as it requires no maintenance or hardware upgrades. Serverless databases also allow for automated repair procedures, hourly backups, snapshot storage, and unparalleled data access control. With the ability to automatically provision and deprovision storage capacity, developers can optimize their systems for performance, speed, and cost savings. Additionally, serverless designs enable efficient sharing of resources through microservices architecture, multi-tenancy, elasticity, and auto-scaling. This results in significant reductions in total cost of ownership compared to non-serverless database workloads.

Dataworkz, a company based in Milpitas, California, aims to simplify AI-driven decision making by offering an easy-to-understand user interface for its no-code, high-performance cloud service that unifies data, transformations, and AI for business users. The platform is designed to enable the creation of AI-based applications and focuses on addressing challenges related to preparing data for machine learning initiatives, such as obtaining availability of data to run algorithms against and managing software integration problems due to multiple teams and applications involved in bringing in data and combining it all at scale. Dataworkz's solution involves unifying data processing and AI into a single experience offered as a service, allowing business end users to go from understanding the data to transforming it and running machine learning algorithms on it within one platform. The company has multiple production deployments with customers in high tech, legal services, life science, and more, focusing initially on sales and marketing applications. Dataworkz uses Apache Cassandra for its multi-modal persistence layer that includes Cassandra, relational, and graph databases, and has migrated to DataStax Astra DB for a column store due to its ease of setup, strong security features, global availability, scalability, and consumption-based pricing model. The company is excited about democratizing access to enterprise AI and data-driven decision making tools, allowing business users to run data transformations in a secure manner without needing to go through the IT team for each new request or modification.

This post introduces NFTs (Non-Fungible Tokens) and explains how to load them into DataStax Astra DB, a multi-cloud database-as-a-service built on Apache Cassandra. It covers the definition of NFTs, their types of smart contracts, and the process of using the Astra DB data loader to store every NFT ever minted in Astra DB. The post also discusses how NFTs are purchased and sold using cryptocurrencies like Bitcoin or Ether via auction house websites like OpenSea or Rarible. It further explains the concept of NFT smart contracts, their programming within blockchain, and the two main standards (ERC-721 and ERC-1155) that guide developers in interacting with NFTs. The post concludes by demonstrating how to pull NFT transfers, grab metadata, store it in Astra DB, and search for an NFT within the database.

This fourth part of a six-part series on Apache Cassandra introduces Storage-Attached Indexing (SAI) features on DataStax Astra DB, a managed Cassandra-as-a-database-service on the cloud. The text discusses indexing in both relational and Cassandra databases, explaining how secondary indexes work compared to distributed indexing like tables and materialized views. It also covers the limitations of secondary indexing practices and introduces DataStax's SAI as a solution that provides efficient secondary indexing for Astra DB and Cassandra. The text concludes with setting up Astra DB to create a fully-realized Cassandra database and query outside of primary keys through SAI, along with examples of use cases like client data modeling and IoT applications.

Deploying DataStax Astra DB on Google Cloud Platform (GCP) has brought several advantages to the company. Kubernetes, an open-source platform launched and open-sourced by Google, enables them to build their Database-as-a-Service (DBaaS) against a standard cloud platform. GCP's browser-based cloud shell is a quick and easy way for developers to log in to their cloud VMs. Additionally, the platform offers significant sustained use discounts without requiring any commitment from users. These features make GCP an attractive choice for cloud development and deployment.

Ansible, a powerful IT automation engine, is being used by thousands of companies to drive complexity out of their environments and accelerate DevOps initiatives. DataStax Astra DB offers a rich set of APIs, including the DevOps API which makes it easy to script the creation and management of an Astra DB database, roles, security tokens, access lists, and private links for your database. With Ansible, users can leverage APIs as part of their playbook, making it easier to interact with Astra DB’s DevOps APIs and Stargate Data APIs. The use of Ansible helps in automating the process of setting up a free-tier database, creating tables, and managing other operations related to Astra DB.

NoSQLBench Pulsar Driver is an open-source performance testing tool for Apache Pulsar that simulates realistic production workloads. Unlike other tools, it allows users to define workloads in YAML files and supports various types of Pulsar workloads. The driver also enables fine-tuning of behaviors through configuration settings and offers deterministic, realistic, and production-like Pulsar workload simulation. It can detect abnormal message processing errors like message loss, duplication, and out-of-order. NoSQLBench Pulsar Driver is a powerful tool for performance testing, messaging model design, sizing analysis, and cluster deployment evaluation for applications and systems based on Apache Pulsar.

Part 3 of a six-part series on Apache Cassandra discusses benchmarking and stress testing using NoSQLBench, an open-source pluggable testing tool for the NoSQL ecosystem. The tutorial provides hands-on experience with benchmarking and stress testing Cassandra using NoSQLBench, covering its integration with different kinds of workloads and protocols, including CQL support. It also guides users through creating custom workloads for their own applications or databases. The exercises involve running preconfigured scenarios on Katacoda, accessing resources on GitHub, and following step-by-step instructions in a YouTube video.

This tutorial teaches how to build a machine learning model, train it, and turn it into a REST API using Python and Visual Studio Code. The project involves creating a spam detection machine learning model from scratch and deploying it into production. Key technologies used include DataStax Astra DB for automatic connection to Cassandra and Python 3.9 or above. The tutorial guides users through setting up, preparing, and exporting datasets, training the AI Spam Classifier using LSTM model, uploading models and metadata to object storage providers like Linode and Digital Ocean, creating a reusable AI model class, configuring FastAPI app, loading Keras model and predictions, integrating with Astra DB, storing inference data on Cassandra, paginating the Cassandra model, testing the AI as an API through ngrok, and deploying the application into production.

In this tutorial, learn how to web scrape on a schedule by integrating FastAPI with Astra DB. The video tutorial demonstrates using Python and Jupyter Notebook for web scraping, storing data in Cassandra, validating data with pydantic, scheduling tasks with Celery, and handling large datasets with Astra DB. This technique can be used to track price changes, observe trends or historical patterns, and leverage big data analysis. The tutorial also provides resources for learning more about Apache Cassandra and DataStax's Astra DB.