The evolving face of Xoxoday

When Xoxoday(earlier Giftxoxo) started with an idea of experiential gifting in the year early 2013, little was known that it will cover the market so fast and become a market leader in marketplace for Experiences and Activities. Every year that has passed by since then, the company has seen approximately 300% to 400% growth in Experiences consumption. Since an early adopter of the concept in India where every thing is valued against worth of quantity against each penny spent, Xoxoday carved its position well in corporate world there by developing the market for selling Experiences and Activities. Cometh the end of 2016, and this specific category (Experiences) can be found on most of the travel, concierge and event booking platforms.

When the market was gearing up, Xoxoday opened the idea of self-consumption with the launch of new b2c platform. Since then it has seen peak in consumer space and experience is becoming fast trending product for both corporates and regular consumers. The trends suggest that the year 2017 is going to be the year of “Experiences & Activities” and most of the companies operating in spaces like travel, food tech, flight bookings, banking rewards etc will channel the fulfillment either in-house or through third party APIs.

Here at Xoxoday, we believe in bringing disruption to the market and is led by a vibrant technology team.  The idea of getting into hyper local Experience & Activity marketplace means the tech specs has to be supported with all the right weapons. Right from hiring super-engineers to performing super tasks, we have been doing a lot since past one year.

We have brought a lot of changes in server and application architecture in past one year to meet the demands of market and also to plan for next billion users. High availability, Fault tolerance, Scalable and multiple availability zones architecture have helped to provide 99.5% uptime to our clients and consumers. To achieve continuous and quicker deployment, we use lighter Microservices for middleware and dockers for distributed deployment.

We use recommendation engine, Context based filtration and AI to match experiences to user taste/preferences, current location, user past footprint to provide them personalized experience while they browse through our applications.

However shifting from a monolithic to micro-service kind of architecture requires lots of preparation. A sneak peak of our architecture and flow is given below:

To achieve the goals above our architecture has been changing by taking few considerations as:

?  dynamic polyglot runtime

?  fully asynchronous service layer

?  Reactive programming model

The new architecture is a significant departure from previous generic monolithic architecture.

Dynamic JVM languages combined with an asynchronous Java APIs and the reactive programming model have proven to be a powerful combination to enable safe and efficient development of highly concurrent code. This mixed with right technology on backend and frontend, has helped us to look ahead and prepare for the scale we are aiming to achieve.

All the scheduled jobs run on AWS Lambda. Xoxoday has a very efficient cataloging team, which also takes and validates data from crawled dataset, which is also scheduled powered by python and saves data in AWS Dynamo DB. With such huge dataset it is almost impossible to do it manually and hence the team at Xoxoday has developed Drools rule engine which is at a nascent stage but soon to become mainstream. The data is finally presented after transformation and filtration, which is done using Apache Spark on AWS EMR.

The backend is powered by Vertx(vertx.io) and AWS Elastic Cache. The microservices communicate asynchronously by running on Vertx event bus and using REST APIs. At Xoxoday all the entries and browsing is monitored which further helps in enhancing recommendation engine. Combined with a very strong monitoring mechanism using AWS CloudWatch, auditing and logging are maintained using internal microservices.

The team is now focused on intelligent empowerment. Long-time R&D topics like machine learning and artificial intelligence have already penetrated to practical application scenario and team is playing around with POC to be brought in the platform.

All together the space is looking very interesting and right up for disruption.