The European ecommerce giant Zalando is in the middle of a paradigm shift: an evolution from online fashion shop to fashion platform. As part of this evolution, our engineering team has adopted an “API First” approach: publishing APIs that other companies can use to take advantage of our massive amounts of data and build their own applications. Zalando’s primary public API (https://api.zalando.com) is implemented with Java using Spring and RestEasy, and offers programmers access to the web shop. It also allows for basic operations such as searching for articles, categories, filters and brands.
Making our APIs public has presented some very exciting technical challenges: specifically, how do we auto-scale our API as our audience grows? In this talk, we will provide the answers to these and other questions by sharing insights into the architecture-related choices we have made as part of our API First shift. We will describe how we adopted AWS to aggregate data from our catalog and real-time stock data stores to ensure top API performance. We’ll explain how we use Solr and AWS to manage our catalog data store and power a useful, consistent search engine. We’ll talk about how we use Memcached to overcome constraints related to our real-time data store and ensure accuracy of our API data. Finally, we’ll summarize how all of these engineering efforts enabled us to succeed in auto-scaling our public API across multiple regions while keeping our codebase intact. By the end, we hope you’ll have some useful takeaways for managing scalability issues of your own.
