Tarantool is an in-memory computing platform that provides real-time access to massive amounts of data. It is used in high-load scenarios, mostly by large retailers, banks, and telecom companies.

This is a 2021 project. Tarantool was expanding its reach to industries and required several new pages. My job was to write the copy for two landing pages, a case study for the main website, and another landing page for the sister company's AWS-like cloud platform, where Tarantool can run as a service.

I worked directly with the marketing department, which provided a detailed brief for each page. I did some additional research, but I mostly used the information provided by the team. We used Trello to plan the work and move each unit through various stages of production.

Tarantool was used in a sister company's enterprise platform designed to examine business processes and look for discrepancies with regulations and unnecessary approvals.

For this case study, I followed a very straightforward structure:

The client’s profile: what the final service is and how it benefits businesses.

Task and requirements: the absolute must-haves the solution should provide (horizontal scaling, analytics based on thousands of attributes from various systems, real-time data loading, and more).

Possible solutions (PostgreSQL-based solutions, Graph DBMSs, ClickHouse) and why Tarantool was picked instead.

Product's architecture: where exactly in the architecture Tarantool is sitting and what its role is.

Results: the number of events processed without performance decrease, the speed of receiving data from an external source and its turnover, and how long the implementation took from the start to the first pilot project.

Tarantool's design makes it useful for large retail companies with hundreds of brick-and-mortar shops and a centralized online shop where you can place an order to pick it up yourself or have it delivered to your doorstep. Because all inventory is in the memory with fast access, and you don't have to do expensive cache invalidation, you can keep inventory across the entire chain of stores in sync, along with data on region-specific daily discounts.

The landing page outlines several common scenarios in which using Tarantool benefits retail and e-commerce. For offline shops, the page focuses on upselling at checkout, loyalty card processing, and promo processing. For online shops, it promotes SSO, a unified notification system, and, above all, golden records, data marts, and real-time marketing for personalized offers.

Tarantool can be launched on a single node and scales horizontally fairly well. Thus, the landing page suggests that using Tarantool as early as an MVP makes perfect sense. You get the architecture right early on, so your time to market decreases dramatically. At the same time, you get everything you can expect from a relational DBMS—schemas, SQL requests, and ACID transactions.

This is a very short, to-the-point landing page. It didn't survive the website redesign, but it served its purpose.

This is another short and laser-focused landing page. It explains the benefits of using Tarantool as a service on a cloud platform: a quick start with a single click, zero administration cost, fault tolerance, and versatility of use as a standalone database, cache, queue broker, or something else.