|
Market Roundup December 9, 2005 Stacking Up the Tests or Testing Up the Stacks? |
|
Stacking Up the Tests or Testing Up the Stacks?
Red Hat has announced a new service to provide certification and production support for key open source software stacks. The company will offer three new stacks aimed at simplifying and standardizing open source application stacks, so that developers can focus on applications instead of configuring the underlying platform. The new program includes extensive integration and testing of the various components running on the Red Hat Enterprise Linux platform and is initially offering three stacks of open source components. The Web Application Stack is for simple web sites and applications, and includes support for the LAMP components on top of Red Hat Enterprise Linux. Customers may also choose the PostgreSQL database, which is also available as an option with the Web Application Stack. The Java Web Application Stack is for more dynamic web applications, and includes the Web Application Stack plus support for the Apache Tomcat Servlet and JSP container. Included with this stack are support and updates for key Java development libraries and tools: Apache Struts, Apache Axis, Spring, Hibernate, Lucene, Ant, Junit, Jython, Log4J, and key XML libraries. The Enterprise Java Stack includes all of the Java Web Application Stack plus support for a full Java application server based on ObjectWeb’s J2EE-certified JOnAS project. Each Red Hat certified stack includes free access to Red Hat’s Eclipse-based Developer Suite, and will be available as a layered subscription on top of Red Hat Enterprise Linux. The Red Hat certified stacks will be available beginning in Q1 2006 with subscription pricing starting at $599 per server.
This move puts Red Hat in a more competitive position with industry startups SpikeSource and SourceLabs, who are already making a name for themselves when it comes to testing, certification, and support for other open source software. However, since RHEL is the main distribution for Linux open source in North America, this will most likely provide the coattails to give Red Hat an edge in this arena. In addition, there are others who see an opportunity in testing and certification such as Oracle who has also announced its intention to enter the testing and certification of open source applications marketplace. Who said there is no money in OpenSource? Clearly some big players believe there is, but it is not in the software itself; rather in the guarantee that it will work as indicated. This is one of the advantages that the traditional software vendors have had, an implicit guarantee that the underlying framework will work as advertised, or at least a support center to work it through when it doesn’t. With the move towards certification from significant system infrastructure players, we believe this should help application developers remain focused on the value-add of their application as opposed to mundane testing of the underlying framework. We believe this is a good thing. But will Red Hat, who at present has taken a limited stack-specific approach to testing and certification, feel the urge to expand to broader testing such as that of SpikeSource and SourceLabs or will it remain true to its infrastructure roots? Time will tell, but we suspect that many developers will feel the lifting of an implicit load off of their shoulders with this announcement, and may just sleep a little easier tonight.
Novell Identity Management: Third Time’s the Charm
This week, Novell announced Identity Manager 3, its provisioning and identity management solution for security and compliance. According to Novell, Identity Manager 3 provides advanced visual modeling, workflow, and self-service capabilities. The company believes customers will be able to better simplify and control user access, protect data, reduce administrative costs, and comply with corporate and regulatory mandates. Novell plans to do this with automated provisioning which provides consistent service across the network. Novell argues that the key differentiator for Identity Manager 3 is the visual framework for mapping complex identity management infrastructures with automatic documentation, instead of the traditional manual coding and configuration customers have struggled with in the past.
Identity management is a vital and indispensable component of network security, and indeed Novell’s new product sits within its Security and Identity Solutions group as one of a family of solutions that together provide end-to-end security. Novell, like other vendors, has offered earlier versions of identity management, but this edition has new capabilities that should entice users considering Novell or who already use Novell products to take special notice. In this latest release, Novell focuses on new automation features specifically designed to address the impact of governance and regulatory compliance that organizations are encountering, as well as making ongoing administration simpler and more consistent. In particular, automating the provisioning of workflow has been a focus, which means that roles, groups, or individual assignments can be automated; delegation and proxy functions, expiration tracking and escalation, and self-service provisioning are now possible. More importantly, with Identity Manager 3, no coding is required to deploy the product. Novell, who has an edge on the user interface for this class of product, continues to develop a visual framework that is more straightforward than command-line interface requirements for many users. Overall we think this is a solid product and something customers should seriously explore.
Despite a great product, however, Novell may not see the success it deserves. This is for two general reasons. One is that Novell has somewhat lost its way as a company and spent many cycles reorganizing itself in the last couple of years. The amount of energy spent on internal issues has meant fewer resources dedicated to marketing and messaging its products to the market. Ten years ago Novell was ubiquitous (if you’ve been to CeBIT, you know what I mean.) Now, Novell is still a well known company, but most people probably couldn’t say what Novell’s top current products are. We hope that Novell has found a corporate configuration that works, because it has solid products that deserve being strongly marketed. The other issue is a company called Cisco. While Novell is positioning itself as an end-to-end security provider, Cisco is making a lot of noise about what it can do for network security and IT people overwhelmingly identify with this message. The interesting part is that Novell and Cisco aren’t necessarily competitive; in fact, their products can frequently be used together to handle separate issues. But in order to successfully market and deliver an end-to-end security solution, being perceived as a networking company is critical and Cisco clearly owns that crown while Novell’s position is less forthright. If Novell can figure out how to partner with Cisco or find other ways to leverage itself, it may be able to grow to critical mass as a significant player in this market segment. It is an ongoing challenge and we look forward to watching Novell regain its rightful place in the IT universe.
Sun has announced its new Sun Fire T1000 and T2000 servers with CoolThreads technology. The massively-threaded servers allow customers to take advantage of the new CoolThreads technology without having to rewrite applications given Sun’s binary compatibility guarantee on Solaris across all supported systems. The company indicated that the new systems are also the first servers designed from the ground up for Internet workloads and for running current and next-generation web, application, and distributed database systems. Sun also stated that it is working with the open source community to bring Linux and FreeBSD to the UltraSPARC T1 platform as well as removing the barriers to adoption and opening up the UltraSPARC T1 platform to other applications, systems designers, and operating systems through its OpenSPARC initiative. The T1000 is a 1U, 19-inch deep server designed for web and network infrastructures; the T2000 is a 2U, 24.3-inch deep server with extensive redundancy designed for maximum uptime to support application services and web-tier consolidation projects. The T1000 ships six or eight processor cores while the T2000 ships with four, six, or eight processor cores. The T2000 server is available immediately, and the Sun Fire T1000 can be ordered now for delivery in March 2006. The Sun Fire T1000 and T2000 systems start at $2,995.
True to character, Sun is never a company to forgo an opportunity to make some marketplace noise, cause some FUD, and try to rewrite the rules of the game. The extensive set of announcements around UltraSPARC T1, CoolThreads, and OpenSPARC are testimony of this, but also serve to remind us that the Copernican Company still values R&D and binary capability long after much of the competition has become content with less if any adherence to this standard. Besides coming out front for the moment of continuous leapfrogging of myriad industry benchmarks, a perhaps more interesting message beneath is that of energy efficiency. While most of the world is reacting negatively to notable spikes in energy costs, Sun has taken the opportunity to position itself as a more ecologically friendly computing experience that will help data centers reduce energy expense through a reduction in overall heating and cooling costs. Hmm, sounds compelling in today’s climate, but will the coolest (pun intended) new servers be compelling enough to cause an organization to refresh with the cool breeze of a Sun? We are not so sure.
First some positives. Sun has some nifty performance specs, low prices, the new and hip factor (CoolThreads), all that UNIX tattoo and body art legacy, plus Jonathan Schwarz’s pony tail. Some not so positives: its positioning of this product against Industry Standard architectures (x86 and Power), all that UNIX tattoo and body art legacy, and an undying attempt to convince the market that Solaris + SPARC really are the better way for all, even if the market has spoken otherwise. In all fairness, there is some pretty neat stuff here, but the era of technological excellence as the premier purchase consideration factor has faded in all but a smattering of market niches. Low cost + cool technology no longer trumps application availability, and Solaris is no longer the leading development platform for the non-Windows marketplace. We believe the OpenSPARC initiative is de facto recognition of this reality, and not necessarily a bad idea for Sun. Nevertheless, the availability of the applications increasingly determines server platform deployed as for most the underlying hardware is far less important/interesting than the business software that the organization needs to operate. Even if technology religion or predisposition is a substantial factor, just because UltrsSPARC T1 is available does not mean that there will be an automatic competitive displacement of technologies, especially if the customer is not already a Sun or at least UNIX user. The cost of retraining, redeploying workloads, and possibly changing platforms and reacquiring software licenses could trump the energy savings. Of course, absent some kind of systemic competitive advantage (sorry, benchmark superiority is short-lived and does not count), competitive upgrade is even more difficult. So while we are impressed with Sun’s tenacity and continued clever attempts to rewrite the rules and thus redefine the winner, we are not convinced that the targeted market will in fact find CoolThreads as cost effective as Sun may wish, and thus it may come to find itself giving a cool reception, which for Sun, would not be cool at all.