We are a society that lives for choices. In fact, I believe we strive on choice. I recently read that Chipotle has 65,000 menu combinations! Honestly, about 3,000 too many options for me personally but by any means of measure, that’s a lot of choices. By definition the NetApp Cisco FlexPod solution line is all about choice, hence the “Flex” in FlexPod. We are expanding that “Flex” to include Hadoop.
So, what are we announcing?
FlexPod Select Family. The FlexPod Select solutions are designed for dedicated workloads like Big Data, high performance computing (HPC), video services (production, rendering), oil and gas applications and more. The first of these dedicated workloads that FlexPod addresses is Big Data with our Hadoop solutions.
Hadoop is a powerful and essential tool that has extended business analytics to handle big data. However, there are challenges with Hadoop, as there are with any new technology.
Some of these are:
- It can take significant time to become productive with Hadoop
- It is not easy to use or operate with existing staff or skill sets
- Hadoop NameNode reliability: if the NameNode fails, the Hadoop cluster goes down, so addressing this single point of failure is important
- Standard Hadoop requires three copies of the data, requiring additional storage and reducing cluster throughput and performance while straining capacity and resources
Realizing the benefits of Hadoop and analytics in general requires an underlying infrastructure of storage, servers, networking, and connections to handle big data, and which needs to:
- Have a reliable and robust infrastructure for analytical solutions and platforms
- Implement high-performance Hadoop clusters
- Be built on an open partner-based ecosystem to reduce risk of selecting emerging or new tools
- Allow efficient storage for Hadoop clusters
- Scale compute and storage independently and quickly as data grows in volume and ages
Our FlexPod Select with Hadoop configurations include:
- Cisco UCS C-Series servers
- High density, high performance NetApp E-Series storage for data store – E-series connected to Cisco UCS C-Series servers directly (SAS attached)
- Storage connectivity through direct-attach to server
- Data protocol is SAS for interoperability with existing systems (I’d put the benefit here instead of just mentioning it)
- Targeted for data intensive workloads (think customer patterns, video streams, fraud detection) that require Hadoop
- Highly reliable NetApp FAS to protect NameNode store (not for datastore)
- Either Cloudera Enterprise or Hortonworks Data Platform as the Hadoop distribution
And businesses or departments get :
- Enterprise-Class Hadoop:
- Faster time to value:
- High availability:
- Flexible scaling:
A little more on our E-Series Storage Systems and why it’s great for Big Data
- Connectivity with Fibre Channel, InfiniBand, SAS, iSCSI interfaces
- Up to 384 SAS drives, 1.44PB
- Up to 6,000 sustained MBps
From a storage perspective, the E5460 is used in FlexPod Select with Hadoop. Outstanding bandwidth performance is a critical driver for creating bigger, faster solutions. The extreme density of the E-Series 60-drive enclosure saves floor space and helps lower operational costs for capacity-intensive environments. The E5400’s modular flexibility enables custom configurations that can be tailored for any requirement. Add to this bulletproof reliability, availability, and serviceability designed to ensure continuous, high-speed data delivery and you’ve got a Big Data platform that is ready to go.