Virtualization of Enterprise Hadoop* and Spark* Workloads

Virtualization of Enterprise Hadoop* and Spark* Workloads

Virtualization of Enterprise Hadoop* and Spark* Workloads

The BlueData EPIC* software platform offers data center operators the agility and cost performance of virtualized infrastructure for big data, with high manageability and flexibility when integrating into existing data center environments. Even as virtualization has spread throughout the data center, Apache Hadoop continues to be deployed almost exclusively on bare-metal ...physical servers. Processing overhead and I/O latency typically associated with virtualization have prevented big data architects from virtualizing Hadoop implementations. As a result, most Hadoop initiatives have been limited in terms of agility, with infrastructure changes such as provisioning a new server for Hadoop often taking weeks or even months. This infrastructure complexity continues to slow down adoption in enterprise deployments. Apache Spark is a relatively new big data technology, but interest is growing rapidly; many of these same deployment challenges apply to on-premises Spark implementations. The BlueData EPIC (Elastic Private Instant Clusters) software platform addresses these limitations, enabling data center operators to accelerate Hadoop and Spark implementations on Intel® architecture-based servers. Introduction to BlueData EPIC The BlueData EPIC software platform reduces the complexity of big data infrastructure deployments, providing the ability for end users to quickly and easily deploy Hadoop or Spark clusters in a virtualized environment running on Docker containers. These clusters can deliver faster time-to-value for big data, providing the cloud-like experience of Hadoop-as-a-Service or Spark-as-a- Service in their own data centers. The BlueData EPIC platform helps improve hardware utilization, reduces cluster sprawl, and minimizes the need to move data for big data analytics. BlueData EPIC also provides for simplified deployment and administration, while making virtual clusters look and feel like physical clusters for big data analytics. Taking advantage of the power of containers and virtualization, BlueData’s software helps deliver greater agility and cost-efficiency for on-premises big data infrastructure. The benefits of these capabilities include the following: • Business agility. Virtual clusters can be spun up or down in minutes, providing elasticity for capacity spikes, as well as rapid response to emerging business needs. • Data protection. Multiple virtual workloads can co-exist on the same multi-tenant physical cluster, while isolating data on each virtual cluster from the others. • Resource efficiency. Multiple business units and user groups can share physical cluster resources, avoiding the cost and complexity of each having its own big data infrastructure. To meet varying customer needs, the EPIC software platform is available in two editions. EPIC Lite is a community edition of the platform that is available for a single instance, free of charge; it is intended for evaluation purposes and for personal use. EPIC Enterprise is a fully supported, highly scalable commercial edition that is available on a subscription basis for up to hundreds of physical nodes. For a full comparison of the two product editions, see www.bluedata.com/product/comparison. BlueData Enables Virtualization of Enterprise Hadoop* and Spark* Workloads 2 Hadoop-as-a-Service or Spark-as-a-Service in an On-Premises Deployment Model The BlueData EPIC* software platform gives business users the ability to set up self-service virtual Hadoop* or Spark* clusters without having to submit requests for Read the full Virtualization of Enterprise Hadoop* and Spark* Workloads.

Videos relacionados