Hadoop Consulting

Hadoop is open source framework allowing distributed processing on commoditized servers. Fruition utilizes Hadoop for its Google Penalty Checker system and other Internet marketing systems. Out of this experience Fruition is now offering Hadoop consulting.

The power of Hadoop is the ability to utilize fairly simple programming methods to scale out to deliver highly-available applications for enterprises.

Hadoop is good for data that seems random

When you have massive amounts of data that is not easily connected i.e. no foreign key, then Hadoop can provide a method to bring order to what seems random. Further, Hadoop provides the ability to spread out the processing across dozens of servers each with multiple processors. Versus generally having to try and run the processes on one large server.

Hardware Specifications for Hadoop
Fruition typically uses Dell’s C class servers which are built specifically for this purpose. We load them up with 96+ GB of RAM and SSDs. The servers do not need RAID cards or even redundant power. They are really commoditized servers that are easily replaceable. On occasion, we utilize Amazon AWS or Softlayer for Hadoop as well. However, if the project is projected to last more than a few months setting up in a private data center with owned hardware is typically more affordable than a true cloud solution.

Hadoop is not good for data that is easily connected

For data that is easily organized and related, MySQL or SQL are usually very sufficient and perform well. If you have massive databases where the database needs to be split for performance reasons (“sharding”) then there are other options to consider. However, in most instances mysql is practical and easy to use.

Typical Hadoop Projects

Hadoop consulting projects vary from data center setup to programming. Hadoop consulting retainers start out at $50,000. If you are interested in discussing your big data project please contact Fruition today.