Red Hat details next Linux and storage platforms for cloud, big data era

SOURCE: ZDNet

Red Hat is ramping up for the next generation data center by supporting Google’s Open Compute project, software-defined networking advancements such as OpenvSwitch and OpenFlow and making steady advancements in the operating system, virtualization, storage and networking, company executives said at the summit this week.

Red Hat, for example, is optimizing its Linux, storage and virtualization software platforms to hook into Google’s Open Compute project to provide for a more agile and flexible data center, essential for cloud computing. Red Hat Enterprise Linux is ready for certification on Open Compute Hardware.

“Red Hat Linux is the foundation to ensure hardware enablement happens,” said Red Hat CTO Brian Stevens, noting that RHEL-based nodes can become more compute and network savvvy to better participate in elastic storage services. Later this year, Red Hat plans to debut live migration features to its storage platform.

During the weeklong summit, executives detailed how its next generation software will exploit important new virtualization and cloud technologies.

Red Hat Enterprise Linux 7 clusters, for example, will scale up to 200 hosts and expanded virtual PCI bridge support will enable thousands of PCI devices to be connected to each virtual machine, while the VMware limit is 60 per VM.

Additionally, the RHEL 7 NUMA-based balancing solution will offer new AutoNUMA and SchedNUMA features.

KVM, which is integral to the Linux kernel and Red Hat Enterprise Virtualization for Servers and Desktops, is now used by IBM, ebay, Qualcomm and Dutch Cloud. Red Hat claims the open source hypervisor recently achieved a world record IOPS benchmark and the next generation of Linux KVM will support Microsoft’s Hyper-V, sVirt security, QEMU sandboxing and secure wipe of retired VMs.

Storage is another major area of focus as the cloud and big data eras move forward.

On the storage front, Red Hat this week announced general availability of Red Hat Storage 2.0 based on its acquired Glusters technology designed for unstructured data. Version 2.0 offers a unified file and object store (REST API), geo-replication ailback and high availability or NFS and CIFS and OpenStack SWIFT as well as geo-replication and compatibility with more than 50 dual-socket x86 servers.

Red Hat now has in technical preview a new storage management console and Hadoop Plug-in and plans better support for storage resident applications (MapReduce), storage virtualization and file centric storage.

In the next 2.X Storage release, Red Hat plans to offer multimaster geo-replication and the new management console as well as NFS v 4 and Volume Snapshot. Beyond that, Red Hat is working on storage virtualization enhancements such as extending its virtualization platform and oVirt engine to manage storage pools and file centric storage.

For instance, Red Hat intends to leverage its Linux containers to run applications within storage nodes. This will enable highly scalable storage for unstructured data in physical, virtual and cloud deployments, company execs said.

Red Hat also plans to offer file centric storage support in its storage platform to allow for geo replication multimaster asynchronouous replication, write-once ready many capabilities, and multi-tenancy (with encryption security) capabilities should the storage pool be supporting multiple audiences.

Future generations of Red Hat Storage will also offer next generation of storage will also provide better web administration, a powerful search capability, better history and reports and better supportf or Samba and CIFS and possibly SMB 3.

On the OpenStack front, Red Hat is working on a SWIFT interface for Gluster, Gluster image store and replication support for OpenStack and making its OpenShift PaaS work better on OpenStack. The company is also working on a “Quantum” generalization on oVirt to enable interconnectivity between the storage and networking technologies of Red Hat and OpenStack work seamlessly.

Red Hat has Linux-based storage, cloud and middleware components and the big data movement is dominated by Linux deployments.

Big data is “very open source dominated and that puts Red Hat in a good position,” said Scott McClellan, senior dirctor at Red Hat, who notes that the company is optimizing Hadoop to run on Red Hat Grid.