Google’s Cloud Bigtable and Cloud Storage C++ client libraries
The cooperation
Unoperate developed Google’s Cloud Bigtable and Cloud Storage C++ client libraries in close cooperation with Google’s engineers.
This is highly efficient, asynchronous software, which has to be bulletproof, as it underpins a large number of customers using Google Cloud through C++.
Among many achievements were: implementation of asynchronous data streams, automatic, self-regulating mutation batching, performance tuning, parallel uploads and more.
Carried out projects
Development of Cloud Bigtable and Cloud Storage C++, Python and Node.js client libraries
Extending PyTorch with a Cloud Bigtable dataset
Extending Tensorflow with a Cloud Bigtable dataset
HBase mirroring client
A word from Google
I worked with the folks from Unoperate for over two years. They solved some difficult problems for us, including components to maximize throughput from client applications to our cloud services.
We spent some time going through multiple rounds of evaluation and testing. This process gave us confidence in what they’d built — we knew we could put it in front of customers.
