November 2020 Summaries

The Aerospike Standup Vol. 17 is a newsletter written by developers for developers, covering blog posts, community questions, documentation updates, releases, and events from the Aerospike team. It shares updates every other week and can be subscribed to via email. In this issue, it discusses Tie Breaker Functionality for Aerospike Multi-Site Clustering, release notes for various clients and tools, and updates on documentation including new/updated docs for Aerospike Connect for Pulsar (outbound beta), JMS, Kafka, known limitations, ACT test results for SSDs, and asbackup docs with partition scanning options.

The Aerospike Standup is a developer-focused newsletter that compiles various resources such as blog posts, community questions, documentation updates, and release notes from the Aerospike team. It highlights new articles like the tie-breaker functionality for Aerospike multi-site clustering and updates to known limitations and ACT test results for SSDs. The newsletter also provides updates on documentation for asbackup with partition scanning options and lists recent release notes, including the availability of tools and server Docker images for linux/amd64, as well as updates for Go, C, C#, and Java clients. Additionally, there is mention of an update to the monitoring stack, emphasizing the ongoing development and support for Aerospike tools and functionalities.

The text discusses the implementation of multi-site clustering in Aerospike using strong consistency and rack awareness features. It explains how to construct a multi-site database cluster with full data replication across multiple sites, ensuring continuous operation even during network partitions or site failures without any loss of data. The article also presents an alternative arrangement for two data centers and introduces the concept of a tie breaker node in recent server releases. This feature eliminates the need for operator intervention and reduces hardware usage while maintaining full data replication across multiple sites, making it suitable for building fully resilient distributed Aerospike clusters.

Aerospike has announced the release of Aerospike Database 5.3, which introduces Expression-based XDR Filtering and expanded Multi-Site Clustering. The new XDR filtering feature allows records to be filtered with Expressions before being shipped to XDR destinations, reducing network traffic, storage requirements, and egress costs. This can also help achieve GDPR compliance and similar user privacy regulations. Additionally, the maximum RTT latency is now specified by a dynamic configuration parameter 'connect-timeout-ms', allowing for more flexible multi-site clustering configurations. The release also includes minor features such as fetching configuration items from environment variables and programmatic access to record size.

Aerospike Database 5.3 introduces Expression-based XDR Filtering, which allows for dynamic filtering of records shipped to XDR destinations based on Expressions and metadata access. This feature can significantly reduce network traffic, storage requirements, and egress costs by enforcing geographic restrictions, GDPR compliance, and other user privacy regulations. Additionally, Aerospike 5.3 expands Multi-Site Clustering with the introduction of a dynamic configuration parameter 'connect-timeout-ms', allowing for more flexible cluster configurations and improved RTT latency estimates. Minor features include fetching configuration items from environment variables and programmatic access to record size, enabling more efficient cluster management and data analysis.

Aerospike is a high-performance key-value database that does not natively support multi-record transactions (MRT) at the server level due to potential slowdowns and non-linear scaling. However, MRT can be achieved in software using the capabilities offered by Aerospike. The package "multi-record-txn" supports atomic multi-record updates in Aerospike through the use of locks created with the CREATE_ONLY flag. Rollback is accomplished by restoring stored values and releasing locks, while full details can be found in the project README. Although MRT may not be as necessary in key-value databases as in relational databases, there are still use cases that demand it, such as workflow systems. The value of this API is greater when using Aerospike Enterprise and Strong Consistency, which ensures no record divergence during network partitions or process crashes.

This article provides a step-by-step guide on how to set up and run Aerospike on Amazon's Kubernetes Service (EKS). The process involves using various tools such as eksctl, helm, kubectl, and the AWS CLI. The author also discusses the pre-requisites for setting up EKS, including installing these tools and creating an IAM user with appropriate credentials. Once the EKS cluster is set up, Aerospike can be installed using Helm, a package manager for Kubernetes applications. The article concludes by demonstrating how to use Aerospike via benchmarking software and provides instructions on how to clean up resources after use.

Aerospike, a key-value database, offers DRAM economy by maintaining a constant 64 byte footprint for every object regardless of its actual size. This allows efficient management of large numbers of objects with limited memory resources. However, challenges arise when dealing with small objects or specific cloud environments where memory to disk ratios are fixed. Techniques like "blocking" and Aerospike's comprehensive List and Map API can help manage these situations efficiently. By storing multiple smaller objects within larger ones, the memory footprint can be reduced significantly. Bit-masking is used to map from a logical object key to a physical container object key, allowing efficient retrieval of data without transmitting entire physical objects across the network. This approach not only reduces memory usage but also optimizes network capacity utilization, contributing to overall performance and economy in Aerospike's database operations.

Around 1.7 billion people globally are unbanked, with the largest numbers in China and India. This presents a significant opportunity for financial institutions to offer banking products and services that can generate substantial new revenue while providing a crucial social service. Many of the unbanked are low-income citizens who face difficulties cashing checks, paying bills, and making online purchases without access to traditional banking services. By offering mobile banking apps, eliminating obstacles such as minimum fee balances or service fees, and introducing prepaid products, financial institutions can help improve the lives of millions while expanding their customer base. Innovative technologies like artificial intelligence and distributed data platforms are essential for serving the unbanked at scale and providing a seamless customer experience.

The Aerospike Standup Vol. 16 is a newsletter for developers that covers various topics such as blog posts, community questions, documentation updates, releases, and upcoming events. In this issue, automated Aerospike all-flash setup, Aerospike Connect for Presto, probabilistic data structures, multi-site clustering, rack awareness, and server Docker images are discussed. Additionally, the monitoring solutions page has been updated, FAQs have been added, and new/updated documentation is provided. Upcoming events include webinars on financial services institutions' use of data analytics and utilizing AI & data for agile banking & fraud management, as well as the CIO 100 2020 conference.

The Aerospike Standup is a newsletter for developers, providing blog posts, community questions, documentation updates, and release notes from the Aerospike team. New articles cover topics such as automated setup, data structures, and clustering, while updated documentation includes monitoring solutions and FAQs. The newsletter also features release notes for various dates, including October 28 to November 4, and tools like Docker images for Server CE and EE.

The aerospike-ansible project is a collection of Ansible scripts that automate the configuration and deployment of Aerospike clusters, including optional features such as TLS, encryption at rest, strong consistency, Java benchmarking clients, and Prometheus/Grafana monitoring stack. These scripts are designed to work best on AWS but can be adapted for other environments. The repository provides detailed instructions and a video demonstration of the setup process. This solution is intended to help users quickly set up Aerospike clusters for proof-of-concept, development, or production purposes while keeping costs low by allowing easy teardown. Future updates may include additional recipes for operational procedures like rolling upgrades and cluster migration.

The financial services industry is evolving with changing consumer needs, driving the need for data processing at scale to address diverse client requirements. Financial institutions are adopting NoSQL databases to handle massive amounts of unstructured data and improve real-time decision-making capabilities. Key benefits of using NoSQL databases include improved fraud prevention, real-time digital ID validation, personalized offerings, targeted marketing, and pre-trade assistance. Aerospike's Hybrid Memory Architecture™ (HMA) offers a unique approach to data persistence and management, enabling organizations to deliver high-quality services in the rapidly changing financial landscape.