The latest release of Red Hat’s top-of-the-line enterprise Linux solution, Red Hat Enterprise Linux v.4 (RHEL4) brings increased stability, scalability, and security for medium to large-scale commercial deployment environments. Major enhancements in RHEL4 include, the 2.6 kernel, SELinux integration, improved auditing abilities, storage subsystem performance, autofs4, NFSv4, standards compliance (EAL 4+), upward compatibility for RHEL 2.1/3 applications, and numerous desktop improvements.

Installation

The installation package includes five CDs, compared to six in RHEL 3. The fifth CD is only required for additional application installations. A full install (CDs 1-4) of RHEL4 AS (advanced server) went smoothly and took approximately 65 minutes. A generic PC with 1.5Gb RAM, a 1.5Ghz AMD processor and 250Gb IDE disk was used for the installation. RHEL4 was able to detect the Adaptec 2400A RAID controller and mount an existing RAID5 partition without any further configuration.

Enhancements

2.6 Kernel

Red Hat adopted the Linux 2.6 kernel for RHEL4 only after the code had undergone numerous update cycles within the kernel development project, as well as rigorous testing in versions 2 and 3 of Red Hat’s community-supported distribution, Fedora Core.

The 2.6 kernel brings several major technical advances into RHEL4. First, a new I/O scheduler features four different options (NOOP, completely fair queueing (CFQ), deadline, anticipatory) to optimize system performance for different application workloads. These options can be selected at boot time by passing an argument to the kernel. Second, a new process scheduler provides improved scalability in multiprocessor/multicore and hyper-threaded CPU systems, yielding greater performance gains for threaded applications. Also, concurrent data access by multiple sources is much more efficient. Finally, support for upto 32 x86 processors is now possible, a great improvement over RHEL 3, which could only support 8 processors.

Security

One of the most significant security features in RHEL4 is the inclusion of Security-Enhanced Linux (SELinux). SELinux provides a Mandatory Access Control (MAC) environment, where the kernel grants/revokes permission for actions performed by an application, based on a set of policy rules. This approach prevents damage to the entire system due to a single compromised application.

The new auditing feature (”audit”) provides a generalized way to audit both SELinux and nonSELinux events via user-space processes. Auditing system calls is possible (although turned off by default) by loading rules that describe what information should be logged. A basic set of reporting tools is included, easing management of audit logs.

Storage

RHEL4’s default filesystem ext3, offers increased performance. First, space pre-allocation results in faster reads/writes. Second, hash trees are used to implement large directories, for fast directory lookups. Third, dynamic expansion of mounted filesystems is now supported. Finally, improved locking techniques offer better performance on SMP systems.

RHEL4 includes an updated LVM2 (Logical Volume Manager 2) which offers improved support for large storage subsystems. LVM allows unification of multiple storage devices for ease of management. This approach offers maximum flexibility, as storage allocation matches application needs by abstracting the underlying physical storage. LVM2 brings greater stability with a redesigned, more robust implementation featuring transactional metadata updates, read/write snapshots, and improved storage management tools. The Anaconda installation tool now supports configuration of logical volumes during initial installation.

Conclusion

Red Hat Enterprise Linux v.4 delivers a robust scalable open-source platform well suited for a wide range of mission-critical services. The Linux 2.6 kernel offers increased scalability on large SMP systems, more so than previous versions of RHEL. With the latest productivity applications for Linux, RHEL4 is also a strong contender for the corporate desktop. Looking ahead to future releases, better management tools for SELinux policy rules and SELinux documentation, would ease administration of large deployments.

To many, open source software (OSS) lost its geek fringe status long ago. Now OSS is becoming a leading destination of VC investment and is among the foremost generators of software and ideas. Supporters foresee the day when their approach will replace proprietary software as the leading platform for automation. But open source faces rising hurdles we can ignore only at our own peril.

Beyond ideology

Orthodox GNU, a kind of paleo-OSS, promises a progressive way to organize the software economy through a social contract among developers and users. Yet many advocates of software freedom are no longer driven by the purity of values that founded the movement. An impoverishment of ideals has followed the ironic comity of self-interests that engineered the boardroom coup of free software by the feel-good folks of Silicon Valley. The neologists of the “Open Source” moniker argue that overemphasis on ideology long ago reached the point of zero growth.

Pursuing more practical tactics, OSS mirrors the values and processes that built up the Internet. Like its spiritual teacher, OSS is driven by a global consensus economy which trades influence and advantages on a playground of simultaneous cooperation and competition. This more refined game of co-opetition trumps the knee-jerk gladiator marketing of some proprietary players whose goal is to smash opponents at any cost.

Riding the advantages of its more refined economic model, OSS demonstrated extremely rapid growth early in its life-cycle. Yet OSS can stagnate again because its free-for-all model is often riddled with inefficiencies and, furthermore, lacks the power of principles to drive truly collective collaboration.

Tame inefficiency with good leadership

Consider the redundancies of oocalc vs gnumeric, KDE vs GNOME, Red Hat vs Novell, Debian vs Ubuntu, or the seemingly unbounded license proliferation at OSI. Senseless duplication and competition among products and projects helps no one. The primary challenge lies in taming inefficiencies while still providing differentiation and value. Specialization must be harmonized through a progressive and larger vision as well as good leadership.

Platform neutral organizations like the industry-backed Free Standards Group (FSG), the increasingly global Open Source Development Lab (OSDL), and a newly sensible-about-licenses Open Source Initiative (OSI) together offer the real prospect to provide this vision and leadership.

Be practical but don’t forget the goals

The advantages to OSS of adopting more flexible and pragmatic values are widely acknowledged. At the same time, certain disadvantages of watered-down ideals can be offset by encouraging active collaboration through shared goals instead of dependency on rigid ideology. On the part of industry, cooperation via shared goals is facilitated by striving to meet common customer values and interests. On the part of governments, throughout the developed and developing worlds, collaboration is encouraged by articulating and pragmatically pursuing common societal goals rather than those of monopolists and vested interests. A great example of addressing broad social goals to help bridge the digital divide using techniques of global collaboration is the recent promotion and distribution of localized language versions of OpenOffice.org and other open source tools by India’s Centre for Development of Advanced Computing (C-DAC).

Contribute

As beneficiaries in the global OSS ecosystem begin to come online, simply copying and using is no longer productive. Participants, especially big industry and government, must contribute to generating new technologies and approaches.

Substantive contribution is far harder, but vastly more rewarding, than simply consuming technologies supplied by either proprietary or open source providers. Because OSS will stagnate if collaboration fails at any link in the chain, beneficiaries of OSS must give back to the common pool in order to scale up OSS’s Beowulf-like engine of collaboration. Fortunately, as OSS proliferates, more users are becoming contributors, so grand projects like Spain’s LinEX distribution is both a large OSS consumer (80,000 desktops) and now a valuable contributor to the open source commons.

We often assume that proprietary interests are the main enemies of OSS, but we must also be aware of the challenges we create for ourselves — redundancy and inefficiency, losing sight of shared goals and failure to contribute back to the commons. In the end, OSS must guard against becoming its own worst enemy.

The second annual Open Source Business Conference (OSBC) in San Francisco brought together key leaders of OSS, as well as players from the opposition — under one roof, to examine the latest trends and business practices in open source. OSBC offered a unique opportunity to discuss, absorb and influence developments that impact adoption and mainstreaming of OSS across the globe. VCs, industry leaders, managers and lawyers all took up unique vantage points along the conference line “Innovation in Open Source”. Vendors such as IBM, CA, Sun, and even Microsoft, advocated their own best practices. Some, louder than others.

“OSS means freedom, OSS means business. OSS means war!”

Amid the rattling of sabres, the license bashing and the competitive business posturing of many of the key players, one of the most important voices on the modern software battlefield was Lawrence Lessig whose keynote lecture was on “Clearing the air about Open Source”. Lessig was electrifying. He warned that after the change of US government in 2000, the new administration invalidated ten years of antitrust litigation by simply changing the rules — which in turn helped dismiss the anti-trust proceedings against Microsoft. This action in itself strengthened the monopolist’s strategic mission to acquire an arsenal of software patents as fuel in its war of hegemony to control competitors as well as technology innovation, and business in general. So far, Microsoft has mostly used its patent war chest to defend its product space and market share, but that could easily change as their needs do. Lessig observed that monopolists would spend their current net worth to defend their turfs and crowns. And Microsoft is no different.

Lessig challenged the industry and innovators to think about the consequences of Microsoft succeeding in its patent warfare, and thereby stifling Open Source, innovation and freedom. The US software industry desperately needs to push legal reform in order to maintain free competition and avoid crippling patent warfare and frivolous litigation from monopolists. The danger is being felt all across the world — in the US as well as in proxy battlegrounds like Brazil and India. Developing countries are being pressured constantly by Microsoft and its supporters to believe that free software is more expensive than its own products and services. Brazil, a key battlefield in this war, has officially countered that whatever the expense, free software supports its national interests in helping build a local infrastructure, promoting local innovation and nurturing its fledgling economy. Lessig noted that Brazil believes strongly in building a free culture — free to innovate, free for expression of creativity and free to grow. He warned that in a world dominated by lobbyists and lawyers, the open source community and industry must follow the pattern of Brazil. The OSS community has no choice but to fight to preserve its rights to create open software, to innovate and to grow.

In the end, Lessig advised the open source industry to follow a “republican-like centrist” way of controlling software patents. Every software patent ought to be put through a rigorous economic examination and proved to do more good than harm, and be discarded if demonstrated otherwise. Presumably few, if any, software patents would survive under the bright light of objective scrutiny.

OSBC represented a small skirmish in the war over whose software will dominate the information age. Much larger battles, with risks of far greater casualties, loom, especially in the developing world. For example, in India, as elsewhere, monopolists are in pitched battle against competition and free markets. Global pressure tactics are being used on a local but still lethal scale. As a result, India recently capitulated for a limited time to the WTO, Microsoft and other special interest lobbies, to allow software patents to be recognized and enforced. But as IP monopolists with deep pockets continue their crusades, how long can countries like India defend themselves in a war of unequals? How long will the rights to innovate, promote local businesses and build infrastructure survive? As Lessig eloquently argues, the only way to stop patent colonialism, is to reform the local legal process to promote competition. And continue to adopt, innovate and grow healthily with open source software.