Qubole’s co-founder and CEO, Ashish Thusoo recently addressed attendees at Data Driven NYC #26 in April. He focused his remarks on the benefits of using big data in the cloud as opposed to an on-premise infrastructure.
Thusoo brings a unique perspective to the cloud vs. on-premise debate having worked on both sides of the issue — with Facebook he worked with large, on-premise big data infrastructures, and with Qubole and other companies he’s worked with big data in the cloud.
In the presentation, which was followed by a Q-and-A, Thusoo highlighted three important differences between big data in the cloud and big data on-premise including:
For each of his points, Thusoo highlighted reasons why cloud computing outdid on-premise computing.
Companies can get started much quicker with big data in the cloud than if they were using on-premise infrastructure — instead of taking weeks or months to start, the cloud does it in days. Additionally, the startup cost is significantly reduced because there’s no costly infrastructure to install.
Big data in the cloud also reduces risk by transferring a large part of that risk to the service provider. Individual companies need not concern themselves with system failures, server issues, new servers, etc. All the system risk is with the cloud provider.
Finally, big data in the cloud gives companies much more flexibility. If companies need 100 nodes they can get them on demand. If a company’s needs are changing from week to week, they can easily change their storage and analytical capabilities too. With on-premise big data if there’s any type of system failure it’s difficult to continue running the processes at regular speeds, but those types of issues don’t exist with big data in the cloud.
Data-Driven NYC gathers together monthly to discuss all things big data. Watch the full presentation below.