CA World: Storage a Major Focus, Part 2

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More .

The Web Services Spin

A core element of the move to on-demand computing is web services and this topic came in for discussion again and again at CA World. From a traditional systems management perspective, some view web services as a new layer on top of the conventional management layer that simply maps devices and software infrastructure to the services constructed out of the many components lying within the seven levels of the OSI stack. However, that concept came under attack, with IDC preferring to look upon web services in terms of “Service Oriented Management” (SOM).

According to IDC analyst Rob HailStone, SOM makes it possible “to manage directly at the service level without being concerned about the supporting infrastructure and device technology. By intercepting web service requests, considerable power is reserved for management tools that implement a rules engine approach to determine actions that result from web services events.”

SOM tools can verify identities making requests, reroute requests to alternative service providers, load balance servers, alter messages, and analyze performance and utilization. One speaker gave the example of a company called JETS that offers comprehensive web services-based travel information. JETS own web service actually relies on web services from a wide range of other vendors and corporations to keep track of rapidly changing flight information, credit card data and hotel reservations. Instead of having to access a multitude of servers and deal with an array of differing corporate security systems, JETS web service monitors everything at the message, or Simple Object Access protocol (SOAP), level.

In order for web services to add value, though, it has to move away from the concept of putting agents on servers or gaining trusted access status for secured servers.

“Web service security works best by implementing encryption and signature authorization technologies at the XML message level,” said Dmitri Tcherevik, CA’s web services director. “As messages flow, you look into the message header to pull information so you know what it is, add data, authorize, or route to another web service.”

CA launched several new products to take advantage of these capabilities:

Unicenter Web Services Distributed Management (WSDM)

WSDM automatically discovers and monitors web services across the enterprise. This enables IT to closely track a full range of performance indicators and rapidly respond to service interruptions. This reduces the risk of business interruption and inadequate performance from web services.

Unicenter Management for WebSphere 3.5, WebLogic 3.5, and .NET Framework 3.0

These are three separate products for different platforms that monitor web services from J2EE and .NET application servers.

eTrust Directory 4.1

This is the first enterprise-ready UDDI (Universal Description, Discovery and Integration, an XML specification) implementation suitable for large scale deployment of web services. It can store, replicate and distribute over 100 million entries of web service data.

Page 2: Linux Lovefest

Linux Lovefest

Most sessions at CA World had 30 to 50 attendees. Linux Day kicked off with over 200 people seated and many more standing at the back. The remaining Linux sessions were also better attended overall than those for other subjects. One panel raised the issue of the future of UNIX in light of the explosive growth of Linux. Most agreed that UNIX would gradually fade away. According to IDC’s forecast for worldwide paid server OS shipments, by 2006, UNIX will be down to a few percent, with Windows 2003 server and Linux predominating.

“Each UNIX vendor seems to be offering a UNIX support environment today while offering their customers a path to move to Linux,” said Torvalds.

While there was plenty of drum beating for open source, several of the speakers focused on what Linux needed to make it truly ready for the enterprise. The consensus was that lack of desktop maturity and support seemed to be the most immediate barrier to wider scale adoption.

“Linux is still missing pieces to make it a completely enterprise ready desktop platform, but it is getting there,” said Juergen Geck, CTO of SuSE Linux. “It is also missing pieces on the systems management and hardware support side.”

Desktop progress, though, is being made. The City of Munich, for example, is implementing 14,000 Linux clients across the city. Other government agencies in Europe are following suit.

IDC’s Dan Kusnetsky, however, poured icy water on some overexcited Penguins by making it clear that Microsoft isn’t about to roll over and die. If anything, he said, Microsoft could get even stronger.

“In 2002, the combined revenues for all Linux systems (server, desktop, embedded) reached $86 million compared to about $8 billion for Microsoft,” said Kusnetsky. “On the desktop side, Linux currently accounts for 1/10th of one percent of the total desktop revenues.”

Meanwhile, CA continues to step up its support for Linux. It has ported all its storage and other core products to Linux and is promising that its complete range will eventually be available in a Linux version. Like IBM, HP and even Sun, CA is now investing heavily in the future success of Linux.

Virtually Not Yet

For all the show’s talk about virtualization and on-demand, the reality is that it will be many years before real world reality matches the hype. This latest industry buzzword, however, seems more on a par with the early days of the Internet mania than more dubious fad’s such as ASP. That said, on-demand storage and the on demand data center are still a ways off.

“Right now, data centers are experiencing the early days of on demand in terms of automation and virtualization of servers and storage,” said Meta Group’s Fenengul. “But it will be another five years before everything comes together at the policy level and as much as ten years before we arrive at something close to the utility computing concept – a shared service-centric variable cost model that is customer flexible.”

Missed Part 1? Click here.