Red Hat global preferences

Red Hat Blog

Red Hat Enterprise Linux 6: A Technical Look at Red Hat’s Defining New Operating Platform

 

 

Download this video:[Ogg Theora]

Today, we delivered Red Hat Enterprise Linux 6 to the market, and I’m thrilled to have this opportunity to share with you, on behalf of the whole Red Hat Enterprise Linux team, exciting technical news about why Red Hat Enterprise 6 is unquestionably the most comprehensive Red Hat Enterprise Linux release in Red Hat’s history. I’ll take you through a look at our key objectives, development approach, and high-level feature themes. I certainly hope that you’ll soon see why we are so excited to deliver well over three years’ worth of customer-focused product innovation into the hands of our customers for immediate deployment in advancing datacenter efficiency.

Before diving into the feature detail, let’s start with the development model. When establishing objectives for a major release, we combine inputs of technical possibilities from our creative development team with the voice of our customers’ input regarding practical business needs. We balance that off with an experienced assessment of where we can lead in innovation with the collaborative open source community. Hand-selecting technologies in a responsible manner that can be supported for a long-term, 10-year lifecycle is not an easy balancing act.

An apt analogy is to picture the broader open source landscape as a universe of separate solar systems and galaxies. Each spinning and revolving largely independently. Similarly, the broad diversity of open source projects each are largely autonomous – in that they are not well-integrated and each are at different points in time in different lifecycle phases. At one point in time, some projects are in their stable phase, at other times they are undergoing turbulent innovation. Due to the fact that Red Hat is one of the leading industry contributors to major Linux open source initiatives, we are well positioned to drive customer-centric innovation. Red Hat has the experience to identify enterprise product ready project versions and address the massive integration challenges to yield an effective enterprise offering.

In my development team we feel a tremendous sense of pride in that we have the privilege of being at a confluence point of technology innovation and customer need fulfillment. This gives us a full lifecycle of gratification in that we get to build it, test it, harden it and learn from this where the next iteration of technology advancement may flow. Red Hat is well positioned to provide the 10-year support lifecycle for Red Hat Enterprise Linux 6 – nobody can support something better than the team that leads the technology building and integration. From our perspective, there’s no greater praise than to see the releases harnessed by customers – that’s a win-win situation.

Now let’s look more closely at how we put this theory into practice by touching upon a few of the high-level themes and representative features of Red Hat Enterprise Linux 6. In the interest of brevity it is not possible to comprehensively cover the literally several thousand enhancements comprising Red Hat Enterprise Linux 6. And so, we will touch on the following themes:

  • Scalability
  • Datacenter operational flexibility
  • Reliability
  • Open

Scalability

Today’s commodity computers, introduced in the past year, rival the scale (in terms of memory, processor counts) of what was previously the exclusive domain of proprietary high-end UNIX offerings. For example 5U rackmount systems have been shipping since the spring of this year containing 64 processors and 2TB of memory.

Effectively harnessing computing resources on the scale of not only today’s hardware, but also anticipating successive generations ahead has been a key Red Hat Enterprise Linux 6 objective.

As a result, in many workloads Red Hat Enterprise Linux 6 achieves the following performance advancements:

  • up to 2X improvement in network rates
  • 2X to 5X improvement in multiuser filesystem workloads
  • Virtualization I/O enhancements allowing increased consolidation (more guests per host) while at the same time reducing I/O overhead significantly in comparison to bare metal. Thereby making Red Hat Enterprise Linux 6 a virtualization deployment platform well-suited for I/O-intensive workloads such as database, mail and file serving. Features such as these virtualization I/O scalability enhancements help Red Hat Enterprise Linux stand out with industry-leading performance.

Datacenter Operational Flexibility

Today’s larger computers generally are not deployed as single purpose. Rather, consolidating workloads – whether bare metal, virtualized or cloud deployments – necessitate fine-grained control of both physical resources and security isolation. This is where we believe the elegant model of an integrated virtualization platform – based on KVM – really shines in Red Hat Enterprise Linux 6. Because virtualized guests are fundamentally implemented as secure processes, that enables generic Linux feature enhancements to directly augment virtualization use cases. This consistency obviates the inefficiencies of implementing everything twice (once for bare metal, again for virtualization). A few key examples will help to illustrate the point.

  • Control groups – enables the system administrator to control resource consumption – for network & disk I/O, memory consumption and CPU utilization. Some interesting practical use cases include setting policy on which virtual guests have higher priority.
  • Svirt – refers to SELinux-based security enhancements for virtualization – enabling policy to constrain each virtualized guest’s ability to access resources like files, network ports and applications. This forms a two-layer check and balance system whereby in a multi-tenancy environment if one guest were able to exploit a vulnerability in the virtualization layer, the enhanced policy is designed to block that guest from accessing resources of other virtualized guests or the host platform.
  • More Efficient IT – from a power consumption perspective, the most efficient CPU is the one that is powered off – especially important for large systems – ie 64 CPUs. Red Hat Enterprise Linux 6 combines a new system scheduler with more intimate knowledge of the low-level hardware topology to yield improved energy efficiency. We have measured a 25 percent reduction in power consumption versus Red Hat Enterprise Linux 5 for a lightly loaded system – by intelligently placing underutilized CPUs (and other peripherals) into low power states.



Reliability

Red Hat Enterprise Linux 6 ushers in a new wave of reliability focused features that in the past typically exemplified very high-end niche proprietary offerings. These advancements support higher levels of reliability for commodity-level operational characteristics. A few examples:

  • Cooperative efforts with our hardware partners affording enhanced resilience and isolation of hardware failures. For example, fine-grained error reporting allowing us to mark faulty memory pages as unusable, hardware based memory mirroring and failing peripheral isolation.
  • Integration of hardware-based data checksumming at the storage level for improved end-to-end data integrity. This feature has been utilized by the filesystem and volume management layers.
  • A new automated bug reporting utility that captures the state of applications and system service crashes and can aggregate this information either centrally on premise or to automatically log incidents with Red Hat support.

Reliability isn’t just features – rather its core to our development experience. Its about putting the ‘E’ (enterprise) in Red Hat Enterprise Linux. Reliability manifests in many ways, such as:

  • Taking a long-term support horizon. Our approach focuses on integrating mature technologies that we are responsibly prepared to stand behind for the 10-year lifecycle. Ironically, it is not uncommon to see our competitors deliver technology before us that Red Hat has helped develop – in cases where we know it is not product-ready.
  • Testing / Hardening. Red Hat Enterprise Linux 6 beta shipped in the spring of this year. Over the 6-month beta period, the release has been relentlessly beat on by our internal automated regression test matrix (incrementally built from 10+ years of product experience), external hardware/software partners and engaged customers. This productization phase ironed out a wide range of issues. To our development team, our reputation is founded on reliability first, feature advancement second.



Open

Open-ness is a core value here at Red Hat that manifests at many levels:

  • Open source collaborative development – Red Hat Enterprise Linux has had a significant influence on the IT landscape by both contributing to and harnessing the potential of the open source development model. Red Hat is a major contributor to open source, yet we are by no means alone. The collaborative efforts of our hardware and software partners (with many of whom we have 10+ years of successful history), engaged customers, and participants worldwide have enabled Linux to surpass the capabilities of any single proprietary UNIX company. In short, no one individual (or company) is stronger alone than when we combine forces together.
    • A great Red Hat Enterprise Linux 6-based representative example is KVM-based virtualization scalability. There we worked closely with component providers and several peripheral vendors to optimize hardware I/O accelerators. We also worked with OEMs to test and harden scalability features on large configurations.
    • Another outstanding large-scale example is the kernel provided in Red Hat Enterprise Linux 6. The kernel is our low-level software that controls hardware resources, security primitives and process scheduling (among many other things). The Red Hat Enterprise Linux 6 kernel is based on the upstream 2.6.32 kernel (of which Red Hat is independently recognized as the leading contributor). Referring back to an initial example of open source being a galaxy of independent projects – the kernel is a world of its own. Within the walls of the kernel, there were many enterprise-ready features developed with leadership from Red Hat that were not integrated upstream at the time of the 2.6.32 based snapshot. Many of these key features have been integrated both into the Red Hat Enterprise Linux 6 kernel and incorporated upstream. As a result the Red Hat Enterprise Linux 6 kernel is a hybrid of 2.6.32 and several successive upstream feature versions. Examples include many KVM-based virtualization enhancements, security augmentation and hardware enablement for the latest systems. This illustrates that the construction of a leading enterprise Linux platform is vastly more than just grabbing upstream packages.
  • Open customer choice – Over the years, many in the IT industry have experienced the effects of proprietary vendor lock-in. Red Hat Enterprise Linux helps to remove barriers to interoperability, increase operational flexibility and allow customers the freedom to choose. This customer choice is not restricted to hardware – due to the fact that Red Hat Enterprise Linux, in addition to hardware certification, also has a broad ecosystem of certified software application partners. This choice and flexibility extends to a range of deployment possibilities. By providing a consistent application deployment platform – ranging from bare metal, to virtualization to cloud – Red Hat is well-positioned to offer customers increased flexibility and help them avoid lock-in.
  • A Red Hat Enterprise Linux 6-based example is libvirt – this is a Red Hat initiative delivering a high-level management interface to virtualization. This abstraction layer is designed to insulate customers from system specific quirks in configuration and management.
  • Deltacloud – Just as Red Hat has been a catalyst for choice at the hardware level, we are offering choice in the cloud industry. Deltacloud provides an abstraction layer aimed at insulating customers from lock-in at the cloud provider layer. Red Hat Enterprise Linux 6 contains what will be a growing foundation of Deltacloud platform enablers.

Conclusion

Red Hat Enterprise Linux has been in the market for nearly a decade, and has been a truly transformative force in the IT industry. It has been an exciting experience for the Red Hat Enterprise Linux team to be a driving force of innovation. Through our efforts, our partners’ cooperation and customer experience, Red Hat Enterprise Linux has become the trusted operating platform of choice among some of the most demanding IT customers, such as Fortune 500 companies NYSE, Amazon, and Salesforce.com. Red Hat Enterprise Linux 6 continues the tradition of providing a stable platform suitable for long-term deployment while safely incorporating new technologies for physical systems, virtualization and cloud.

Over the past three years, we have listened to our customers, and we have responded working side-by-side with both customers and hardware/software partners to deliver the next generation of industry-leading Linux operating system platform. The result: a compelling new platform providing scalability, operational flexibility, reliability and open-ness.

It’s been great to have this opportunity to share with you the enthusiasm and excitement of the Red Hat Enterprise Linux development and test team’s efforts culminating in Red Hat Enterprise Linux 6. We’ve only had the time here to skim a subset of the broader feature set ranging from desktop to datacenter. We certainly hope you quickly become as thrilled about Red Hat Enterprise Linux 6 as we are.