Blogs

Businesses are looking to DataOps to efficiently get the right data, to the right place, at the right time in order to monetize their data and become a leader in today’s digital economy. Containers and Converged Infrastructure are key enables for Data Ops.

Containers provides a DataOps advantage by helping to automate the process-oriented methodology, used by analytic and data teams, to improve the quality and reduce the cycle time of data analytics. You can deploy a machine learning model in a container in a repeatable way. When demand increases, you can scale to another container and load balance as required with Improved agility and faster response times. The main benefit of a container is that it consists of an entire runtime environment. An application, plus all its dependencies, libraries and other binaries, and configuration files needed to run it, are bundled into one package; it can run reliably when moved from one computing environment to another. A container can be moved from a developer's laptop to a test environment, from a staging environment, into production from a physical machine in a data center to a virtual machine in a private or public cloud. You no longer have to worry about differences in OS distributions and underlying infrastructure. That is all abstracted away.

However, a key element in the successful deployment of a container platform is having a robust and flexible infrastructure that can meet the wide variety of requirements in a highly dynamic environment. Hitachi Vantara has a long history of providing highly available and high-performance infrastructures, from end to end complete converged infrastructures with our UCP CI (Converged Infrastructure) platform to a build-your- own approach using Hitachi Virtual Storage Platforms (VSP). Some specific challenges to providing an infrastructure for container platforms are:

• Persistent Enterprise Storage – Data is core for any application. While containers are stateless, numerous applications like data bases and Continuous Integration/Continuous Delivery (CI/CD) pipelines require data persistency at every level. Hitachi provides a plugin to enable stateful applications to persist and maintain its data after the life cycle of the container has ended. Hitachi Storage Plug-in for Containers is based on the industry standard Container Storage Interface (CSI) driver and enables container orchestrators such as Docker Swarm and Kubernetes to automatically orchestrate storage tasks between hosts in a cluster. Hitachi’s Storage Plug-in enables the advanced enterprise storage capabilities of Hitachi VSP G and F series. These storage platforms provide features such as unmatched NVMe performance and scalability, automation, high availability with no data loss guarantee, seamless replication, remote management, and AI Ops for continuous insights across IT operations. 

• Computing Platform – With a wide range of applications that are stateful or stateless, a wide range of flexible computing platforms is also necessary to match memory requirements as well as CPU requirements. Depending on the deployment purposes and wide range of personas involved, different computing technology might be necessary from bare metal hosts to virtual hosts. The type of computing technology is also a consideration when licensing costs are considered. Hitachi Vantara provides different computing options from the 1U dual socket Hitachi Advanced Server DS120 to the 2U quad socket Hitachi Advanced Server DS240. Hitachi Vantara’s compute platforms also provide both bare metal host deployment options and VMware vSphere deployment to fit the vast dynamic computing needs a container platform demands. Hitachi servers also include options for NVMe caching and GPU processor-based options to enhance compute performance.

• Network connectivity – As with any infrastructure a reliable network infrastructure is needed to provide the network connectivity to allow enough bandwidth and security for north-bound and south-bound traffic. Hitachi UCP addresses the network connectivity using spine and leaf design using Cisco Nexus or Arista switches with Arista Extensible Operating System. SAN connectivity is provided by Broadcom Fibre Channel SAN switches.

• Infrastructure Management – having a robust and flexible infrastructure without efficient management will only increase operation in-efficiencies and hinder the ability to increase innovation and agility. Orchestration and automation are the key to operational efficiencies. Hitachi provides UCP Advisor as a single pane of glass management for converged infrastructure, providing automation for compute, network, and storage infrastructure. Hitachi Ops Center is also available with Hitachi VSP for storage management.

Today, all cloud providers provide infrastructure to support containers and container orchestration. However, Every cloud is built differently. Some build their own infrastructure, and each have their own management tools, differentiated functionality, pricing models and policies. Different vendors offer integration and support for different platforms and constantly change the capabilities they have to offer to capture market share. These differences not only cover physical infrastructure components but encompass a diverse range of characteristics, functionality, pricing models and policies. While many users implement containers in public clouds due to the flexibility and agility of cloud infrastructure, concerns about security, vendor lock-in, eggs in one basket, and ROI prompt many users to implement an on-premise or private cloud approach to containers. There is also a demand for reduced latencies that can only be provided on-premise. For these requirements, a Unified Compute Platform (UCP) Converged Infrastructure provides a great solution.

Hitachi Unified Compute Platform (UCP) CI is an optimized and preconfigured converged infrastructure platform. UCP CI offers a broad range of compute and storage components that can be scaled and configured independently to eliminate overprovisioning. With UCP CI, you can optimize your data center to run any container application workload, at any scale combined with the Kubernetes container orchestration platform. The figure below shows a high-level diagram of Kubernetes managing containers and persistent volumes on the UCI CI stack or with Hitachi VSP systems. It can be run on VMware, as bare metal, or as a combination of VMware and bare metal.

For more information on Deployment Options for Kubernetes Container Applications on Unified Compute Platform CI with Hitachi VSP Series see the Architecture guide at this link.

At Hitachi Vantara we empower our customers to realize their DataOps advantage through a unique combination of industry expertise, differentiated server and storage infrastructure and integrated systems.