Hadoop Consulting

Our Goal

If you are looking to use Hadoop and associated technolgies you have began a large investment. You are going to make a large capital investment in the form of server and switching hardware. Your development staff will undertake the learning curve associated with Hadoop and related emmerging technologies.

Hardware Tuning

Production Hadoop Clusters can be anywhere from 5 to 10000 nodes. There are numerouse implications of each hardware decision. Clients can leverage our experience to ensure their cluster has the price to performance ration they need. With data center grade network switching experience we know how to make the underlying network fast and reliable. In depth knowledge and experience of Hadoop gives us insight to help you select server configurations to maximize performance and limit bottlenecks in components like disk and memory.

Cluster Configuration

The Hadoop configuration has many settings critical to cluster performance. As of version 0.19.0 hadoop has over 170 tunable settings! Researching these settings yourself could take days or weeks. We ensure a properly tuned onfiguration for your workload and hardware so you get the best performance.


After setup, we put you in the driver seat by leaving our clients our priorietary Hadoop manual customized to their installation. Our manual includes installation notes, hardware configuration specifications, maintenence checklists, walkthroughs of common administrative functions, and more. This information provides your administrative staff comfort in knowing the cluster is setup and mannaged correctly.

Support and monitoring

While much community focus is on features of Hadoop and related projects our high volume web site and datacenter background had tought us that support, reliability, and performance are what matters most, We make sure your Hadoop cluster is completely monitored with your Network Management System (s). Community initiatives like our Hadoop Cacti integeration, hadoop-cacti-jtg, show the importantance we place on performance and reliability.