June 2023 Summaries

Aerospike's new API calls in version 5.6 make validating cluster synchronization easier. Two common approaches to check if two clusters have exactly the same set of records are comparing record counts and using primary index queries with batch gets. However, both methods have flaws - comparing record counts can only determine if the clusters differ when counts don't match, while primary index queries may miss records in one direction due to data replication. QueryPartitions, a new method introduced by Aerospike, allows traversing one or more partitions in digest order, treating them as very large sorted lists. This enables implementing a cluster comparator that identifies missing records between two clusters by comparing digests and advancing the thread accordingly. The process can be enhanced to also compare record contents if needed. Handling multiple partitions requires concurrency control and selecting start and end partitions, last-update-times, or metadata for comparison. An open-source implementation of this algorithm is available.

Legacy systems in data centres struggle to handle today's mission-critical applications due to their inability to process real-time data at sub-millisecond speeds. This disadvantage can be mitigated through continuous application modernisation, which involves leveraging a NoSQL multi-model data platform that supports JSON document store for efficient handling of diverse and dynamic development patterns. By refactoring applications, optimally storing and querying data, and intelligent data processing at the edge, organisations can ensure their data centres use modern data pipelines to augment mainframes with an intraday operational data store. This approach significantly lowers the Total Cost of Ownership (TCO) while maintaining performance, reliability, and scale.

The economic argument for application modernisation in the data centre is that legacy systems cannot handle today's mission-critical applications and face a dual challenge: losing competitiveness or replacing expensive and time-consuming systems. Application modernisation provides a solution by leveraging real-time NoSQL multi-model data platforms, which allow businesses to manage their legacy applications as an asset to be improved and augmented rather than replaced, keeping costs under control by increasing efficiencies and reducing server usage. The use of JSON document store is particularly effective in this context, as it is ubiquitous, growing in use, and supports real-time data processing and AI/ML algorithms. By modernising legacy architectures through three stages - refactoring applications, optimally storing and querying data, and intelligent data processing at the edge - organisations can lower their Total Cost of Ownership, reduce disruption, and improve business value and scalability. Continuous application modernisation is essential for delivering at scale, improving customer experience, and laying the groundwork for future growth.

Aerospike has launched Aerospike Graph, a massively scalable, high-performance graph database designed for enterprise-scale datasets. The new product addresses the growing demand for an affordable and efficient graph database that offers consistent real-time performance. Aerospike Graph leverages the powerful Aerospike database as its storage layer and introduces the Aerospike Graph Service (AGS) for graph computing, combining Apache TinkerPop and Aerospike Database. The unique architecture enables independent scaling of compute and storage infrastructure to meet application needs while offering an industry-leading total cost of ownership. With support for the Property Graph model and Gremlin query language, developers can easily build real-time OLTP applications in domains like AdTech, Customer 360, and fraud detection.

Aerospike Graph is a massively scalable, high-performance graph database designed for enterprise-scale applications that require real-time performance and affordable cost. It leverages the powerful Aerospike database as its storage layer and introduces a compute layer called Aerospike Graph Service (AGS) that combines Apache TinkerPop and Aerospike Database. The service offers optimized Gremlin steps, compiler optimization strategies, and advanced profiling capabilities to set a new standard for Apache TinkerPop integration. Aerospike Graph is designed to address the needs of both database administrators and developers, offering operational ease and development simplicity while providing an industry-leading total cost of ownership. It empowers developers with support for the Property Graph model, simplifying the modeling of interconnected datasets, and enables seamless integration of Gremlin queries into application programming languages.

The Aerospike Summit 2023 will feature a variety of expert talks on real-time database management systems and their applications in various industries. Key presentations include Ashwin Rao's exploration of AI technologies, Venkat Thamminana's insights into app modernization with Aerospike at Charles Schwab, Subbu Iyer's vision for a real-time data future, Sandeep Nawathe's discussion on Adobe's use of Microsoft Azure and Aerospike for their Adobe Experience Cloud, and Lenley Hensarling's overview of recent and upcoming Aerospike enhancements. The event will take place from June 20-22 with "follow the sun" sessions to accommodate different time zones.

Aerospike, an open-source NoSQL database, was explored during a 1-week internship by student Theodore. He created a functional database and a data visualization program in Python, which took 3 days to complete. To set up the environment, Theodore installed WSL (Windows Subsystem for Linux), Docker, and Aerospike, followed by creating a cluster with 5 nodes and adding an AMS client. The AMS client was used to visualize airline passenger statistics from Data.gov, which showed the number of passengers per airline. The program utilized atomic transactions and plotly for visualization.

The choice of an operational database significantly impacts a business's ability to serve customers, innovate and manage long-term costs. While the initial cost of a database may seem insignificant, it can escalate over time due to various factors such as license costs, infrastructure expenses, staffing requirements, downtime, and competition. Businesses should consider these hidden costs when selecting a database and ensure that its benefits outweigh the total cost of ownership (TCO).

Choosing the right operational database for your business can significantly impact its ability to serve customers, innovate, and manage long-term costs. The selection process should consider factors such as scalability, performance, uptime, and cost of ownership. Hidden costs may arise from database license fees, infrastructure requirements, staff management, downtime, and lost revenue opportunities due to slower response times or outdated features. Businesses should carefully evaluate these aspects when selecting a database solution to ensure it meets their current needs and can adapt to future growth without incurring excessive expenses.