Did you know that not all Hadoop distributions are the same? As Hadoop deployments grow, the architectural differences between Hadoop distributions begin to show dramatic cost differences. These differences can save you 20-50% in terms of total cost of ownership, as we detailed in a previous post.
To make it easier for you to compare distributions and understand the true costs for deploying and running Hadoop, we've developed a simple self-service tool that uses your own data to show how the Hadoop distributions costs stack up. You can easily change the inputs in real time in order to estimate costs across a number of variables in different scenarios. (See the calculator here.)
Key inputs required to start the model include the volume of data, the number of files in initial deployment, and the growth of data (% per year).
Key adjustable assumptions include:
- Discount rate on money
- Number of files per NameNode
- Cost and size of hardware node, including power consumption
- Hadoop full-time equivalent (FTE) costs for administration
- Environmentals (cost of electricity, rack height, cost of floor space)
- Software license/support costs
- Data compression rates
Other adjustable variables in the TCO calculator include:
- Size of drive per node
- # of terabytes/node
- Hardware maintenance costs
- # of nodes managed/FTE
- # of ports per node
- Port types
You can also input information from other Hadoop distributions to see how they compare, both in terms of hard costs (hardware, electricity, floor space), as well as soft costs (labor).
After taking it for a spin or two, you’ll have a much better understanding of the key cost considerations between the MapR Distribution including Apache Hadoop and distributions that rely on the Hadoop Distributed File System (HDFS). The TCO report will include your 3-year total cost of ownership, total hardware costs, environmental costs, and total staffing expenses. You can easily share the results with management via the online TCO Calculator, or other formats such as PDF or PowerPoint.
Have any results you would like to share? Add them in the comments section below.