Running Hadoop on an appliance offers significant benefits, but as Hadoop workloads become more sophisticated, so too must the appliance. That’s exactly why we’re releasing the ‘new’ Teradata Appliance for Hadoop 5. Our new appliance has evolved alongside Hadoop usage scenarios while giving IT organizations more freedom of choice to run diverse workloads. Running Hadoop on an appliance makes more sense than ever before.
If you’re running – or thinking about running – Hadoop on an appliance, you’re not alone. According to an ESG survey reported on by SearchDataCenter.com, 21% of IT organizations are considering dedicated analytics appliances. That’s the same percentage of organizations that are considering public cloud solutions and double those considering a public/private hybrid deployment. What is driving the adoption of Hadoop appliances?
5 Key Benefits of Running Hadoop on an Appliance
Organizations that choose to deploy Hadoop on an appliance versus rolling out their own solution realize five important benefits.
- Hadoop is delivered ready to run.
We’ve heard industry experts say that it can take IT organizations six to eight months to roll out a Hadoop implementation on their own. With a Teradata appliance, we’ve done all the hard work in terms of installing and configuring multiple types of software as well as installing and configuring the operating system, networking and the like. You simply plug it in, and within days you are up and running.
- We’vebuilt high availability into our Hadoop appliances.
The Teradata Vital Infrastructure (TVI) proactively detects and resolves incidents. In fact, up to 72% of all hardware- and software-related incidents are detected and resolved by TVI before the customer even knows about them. We also run BYNET over InfiniBand, which delivers automated network load balancing, automated network failover, redundancy across two active fabrics, and multiple levels of network isolation. These features in Teradata Appliance for Hadoop 5 deliver the high availability IT organizations need in an enterprise-grade solution.
- It is Unified Data Architecture ready.
It’s not enough to just efficiently deploy Hadoop. IT organizations must be able to efficiently deploy Hadoop as a seamless part of an interconnected analytics ecosystem. The UDA-ready Hadoop appliance becomes an integral part of the organization’s larger data fabric, with BYNET over InfiniBand interconnect between Hadoop, the Integrated Data Warehouse and Aster big data analytics, and software integration such as QueryGrid, Viewpoint, TDCH, and Smart Loader.
- Single vendor support.
An appliance replaces the multiple support contracts IT organizations have with their hardware provider, Hadoop vendor, OS vendor, and various utilities, with a single “hand to shake.” If there’s any problem, one phone call puts you in touch with Teradata’s world-class, 24/7, multi-language support for the entire solution stack. IT organizations are seeing increasing value in this benefit as the Hadoop ecosystem has many moving parts associated with it, and single vendor support provides peace of mind.
- Running Hadoop on an appliance lowers your total cost of ownership (TCO)
The cost of a Hadoop appliance includes much more than the hardware the software runs on. There are also costs associated with configuring the network, installing the OS, configuring the disks, installing the Hadoop environment, tuning the Hadoop environment, and testing. The costs for doing all this work internally add up, making the TCO of an appliance even more attractive.
What’s New with Teradata Appliance for Hadoop 5?
In addition to these five benefits, Teradata Appliance for Hadoop 5 delivers freedom of choice to run a variety of workloads. IT organizations now have more options when they run Hadoop on Teradata Appliance 5.
Recognizing that Hadoop workloads are diverse and evolving, Teradata Appliance for Hadoop 5 is available in three flexible configurations, enabling customers to select the configuration that best fits their workloads.
- Performance configuration. For real-time processing and other workloads that require significant CPU, IO, and memory, we offer the performance configuration. This computational intensive configuration enables organizations to run emerging Hadoop workloads such as streaming, Spark, and SQL on Hadoop. With 24 cores, this configuration has more cores per node. It also has 512TB of RAM, 24 storage disks and 1.2TB drives.
- Capacity configuration. The capacity configuration allows IT organizations to drive down the cost per terabyte. It is designed for heavy duty, long-running batch jobs as well as long-term archival and storage. It comes with 128- to 256TB RAM and 4TB disk drives.
- Balance configuration. The balance configuration sits between the performance and capacity configurations, allowing IT organizations to strike the right balance for ETL and analytics jobs. The balance configuration features 24 cores and a 4TB capacity drive.
Learn more about Teradata’s Portfolio for Hadoop.