Ola Ola Cab

Cabin Ola Ola

Who we are </ i>. Rent a car </ i>. Rent a Car <

font color="#ffff00">Taxi India

Ola was established in 2011 by Botavish Aggarwal and Ankit Bhati and is one of the world's biggest rideshare operators. The Ola solution provides a comprehensive, easy-to-use, transparent and fast transport solution for urban transport for clients and drivers on a single mobility technological foundation. At Ola, we focus on making the most of our technologies and developing cutting edge products that are globally applicable.

In 2016, Ola Play, launches the world's first networked carpooling experience, changing the experience of shuttle travel and creating the pace for worldwide innovations in the field. The Ola application allows Ola passengers in over 110 towns and cities with over 1,000,000,000 drivers to communicate via cabins, a rickshaw and a taxi. Powered by a hyper-local paradigm, Ola is dedicated to his quest to build one billion people's lives on the move.

He is co-founder and CEO of Ola, one of the world's biggest carpooling sites and India's most favorite traffic application. He is co-founder and CTO of Ola, one of the world's biggest car-sharing agencies and India's most beloved traffic application. With Ola's rapid scale, he has led breakthrough technology such as map innovation, the world's first networked auto gaming system, Ola Player, and off-line booking that has made Ola available to a large portion of the people.

Cabs Ola about her first two years with Scylla in productions

It' important that our softwaresystems are designed to provide this real-time capability of our businesses. The development of a dispersed system that can meet this demand, however, is a challenging task. A key part of any web-scale system is the data base. In view of the fast growing of our organisation, we wanted our selection of databases to reflect some important features of our work.

The main focus was on supporting high performance, low delay and high uptime ( multi-zone and multi-data center support). Further important demands were that the software is open sourced, requires minimum maintanance and adminstration, has a large tool eco-system and a data base that is used in manufacturing. As many web-based enterprises, we quickly realised that we did not need an ACID-compliant data base for all our use cases.

Correctly designed, we could build many of our frameworks around ultimately consistency, thereby accounting for consistent performance while reaping all the other benefits of an AP-compliant repository like Cassandra. About the beginning of 2016 a new data base, Scylla, attracted our interest. Scylla was promoted as a drop-in substitute for Cassandra, wrote in its mother tongue (C++), and created by the makers of battle-proven softwares such as the KVM hyper-visor, OSCv, and Seastar.

As we realized that Scylla did not yet fully utilize all of Cassandra's capabilities and that it had not yet been tried in combat in production, we were fascinated by the low-level approaches they had taken in creating their databases. As a result of this reassurance, we were optimistic that a long-term wager on Scylla would lead to a profit for us.

A road map has been drawn up to integrate Scylla into our eco-system step by step, on the basis of mission criteria. We used Scylla 1 for our first use in March 2016. Zero in passivity together with a different data base than our main data store. There were only writings done for Scylla. Scylla, as anticipated, has developed quite well in respect of latent and output, with writing optimization through the use of designs.

What surprised us, however, was the robustness of the databank. Next, we began using Scylla 1. This is a constraint of the Cassandra architectural design itself and not specifically for Scylla. In the first cases, we rebuilt our install of Scylla from the ground up.

We used Scylla as a perpetual persistence system in another system, where we could tolerant the risks of losing information due to epidemic nodes. Minimizing the chance of hard drive failure by determining the right Replicator factors, saving across several Availability Zones (AZs), and making sure we could quickly re-generate information.

As a result, we have seen a significant increase in our latent power because Scylla is optimised for running using NTM and avoids an additional net hopping in comparison to EBS. We' ve implemented Scylla for several micro services and plattforms in our company. We use Scylla without a data base buffer in most cases, because by using the right servers node based equipment we could only get our destination data profile (

Mehr zum Thema