Ordinal Science helps clients architect a cloud platform to handle scalable data loads. We design and implement migration plans from legacy platforms to high throughput databases and streaming analytics in colocation data centers or AWS cloud services.
Solve performance bottlenecks with Kafka or Kinesis Streaming, handle IoT volumes with NoSql databases, accelerate analytics with Spark, glean more insight with machine learning.
Our client utilized a solid RDBMS system with 100+ tables in colocation data centers. But rapidly growing databases put a strain on the client’s DevOps team. The burden of tuning the system, maintaining uptime and backups, provisioning new servers to handle the increasing volume disabled new development initiatives.
We recommended a managed RDBMS on Amazon Web Services and proposed a migration plan. Our team modified 500+ stored procedures for efficiencies of the new database engine, adopted data migration tools, conducted full scale testing, then executed migration with only 3 seconds of total downtime.
The migration process was complete in sprints totaling six months in duration. The impact on the client’s customers was minimal during the migration process. And the migration has achieved the primary objective of reducing the database management and maintenance by 75 percent, freeing the DevOps team for other projects.
Identify data pipeline bottlenecks
Assess storage and access strategies
Recommend incremental improvements
Design comprehensive data plan
Design and develop software
Build complete data pipelines
Write machine learning platforms
Develop analytics for big data
Design and execute migrations
Evaluate and clean data
Train machine learning systems
Troubleshoot big data implementation
Kinesis Firehose and Streams
Elastic Map Reduce, Hadoop
EC2, S3, EFS
Apache Spark Analytics,
TensorFlow, Caffe, MLLib,
MySQL, Galera, Mongo
Java, Python, C#
Bitbucket Pipelines, Jenkins