Virtualization Technology News and Information
BlueData Introduces New Innovations for AI and Machine Learning in Hybrid or Multi-Cloud Deployments
BlueData, provider of the leading Big-Data-as-a-Service (BDaaS) software platform, today announced the new summer release for BlueData EPIC. This release builds upon BlueData's innovations in running large-scale distributed analytics and machine learning (ML) workloads on Docker containers, with new functionality to deliver even greater agility and cost savings for enterprise Big Data and AI initiatives. 

Last spring, BlueData introduced support for hybrid cloud environments - leveraging the inherent infrastructure portability and flexibility of Docker containers. This past fall, BlueData delivered a major new release that added deep learning (DL), GPU acceleration, and multi-cloud support to the container-based BlueData EPIC platform. And last month, BlueData announced a new turnkey solution to accelerate AI and ML / DL deployments in the enterprise.

This summer release is the result of collaboration with BlueData's enterprise customers to develop new functionality in each of these areas to support their Big Data and AI initiatives - as they extend well beyond Hadoop and Spark to a range of different ML / DL and data science workloads, and beyond on-premises infrastructure to public cloud and hybrid architectures. These customer-driven innovations provide the agility of containers and elasticity of cloud computing, while ensuring enterprise-class security and reducing costs with automation. Now BlueData customers can benefit from AI-as-a-Service and ML-as-a-Service capability for their enterprise deployments - whether on-premises, in multiple public clouds, or in a hybrid model.

One of the key concepts underpinning this new release is the separation of compute and storage for Big Data and ML / DL workloads. This is a fundamental tenant of the BlueData EPIC architecture, and it allows organizations to deploy multiple containerized compute clusters for different workloads (e.g., Spark, Kafka, Tensorflow) while sharing access to a common data lake. This also enables hybrid and multi-cloud implementations, with the ability to mix and match compute/storage resources (whether on- or off-premises) depending upon the nature of the workload. And it provides the ability to scale and optimize compute resources independent from data storage - delivering greater flexibility, improving resource efficiency, eliminating data duplication, and reducing cost through the reuse of existing storage investments.

Building upon these prior releases and this architectural foundation, the summer release of the BlueData EPIC software platform provides several new features and benefits such as:

  • ML / DL Workloads: This release includes pre-tested, pre-integrated Docker container images for distributed TensorFlow, Caffe2, H2O, BigDL, and SparkMLlib with notebooks - for deployments using GPUs or CPUs, either on-premises or in the cloud. Now enterprises embarking on AI initiatives have the ability to quickly spin up multi-node ML / DL sandbox environments for their data science teams. And if they have an existing data lake, they can easily and securely tap into that data for building and deploying their ML / DL pipelines.
  • Hadoop Cluster Optimization: The summer release includes a new "External Host" innovation that allows containers running on a private non-routable network with BlueData EPIC to act as an extension to their on-premises physical Hadoop clusters. This allows enterprises to easily scale and expand their existing Hadoop clusters - and reduce overall TCO - by adding new containerized compute-only nodes (for Spark, BI / ETL, ML / DL, etc.).
  • Automation for Hybrid and Multi-Cloud: BlueData EPIC now features a new extensible CLI software development kit (SDK) for automated deployments supporting multiple public cloud services. It includes fine-grained automation hooks to enable creation of auto-scaling schemes as well as the use of spot instances in public cloud environments. This same functionality can be used in hybrid cloud deployments.
  • Container Migration: With the new summer release, BlueData EPIC allows for the migration of stateful containers between hosts to support maintenance and disaster recovery scenarios. This feature includes a purpose-built external storage plug-in, and provides the unique ability to move relevant container-specific directories in addition to the data directories - thereby supporting seamless container migration for stateful services.
  • Service Provider Support: BlueData EPIC now has a new RESTful API version that supports sophisticated workflows spanning authentication, authorization, resource allocation, cluster creation, pre/post configurations, storage connectivity, and running analytics jobs. With this new functionality, organizations can map each of their tenants to a unique instance of an Active Directory and/or SAML service - for automated onboarding of multiple lines of business or separate external customers in a service provider model.

"At BlueData, we've been in the fortunate position of helping many of the world's largest and most well-respected enterprises on their Big Data journeys," said Kumar Sreekanti, co-founder and CEO of BlueData. "Now we're working with these and other new customers to advance their AI implementations. In some cases, these organizations are leveraging their existing investments in Big Data technology - and in other cases, they're embarking on new AI initiatives with ML / DL applications. This new release provides these customers with innovative functionality to deliver faster time-to-value, greater agility, and lower TCO for their Big Data and AI deployments."

You can see a demo of the new BlueData EPIC summer release at the DataWorks Summit in San Jose, Calif., this week in booth S5. At the event, you can also learn more about running stateful Big Data applications on containers by joining the session "What's the Hadoop-la about Kubernetes?" with BlueData's vice president of products, Anant Chintamaneni, and director of solutions management, Nanda Vijaydev, at 4:50pm today Tuesday, June 19.

The summer release of BlueData EPIC will be generally available in the third quarter of 2018.

Published Tuesday, June 19, 2018 12:38 PM by David Marshall
Filed under: , ,
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<June 2018>