A Bird's Eye View

Specialized Service Providers Demand End-to-End Management

"Last year we introduced a major operations support and business support systems project to serve as a kind of band leader to pull together multiple disciplines," says Kurt Gastrock, vice president of technology services for Sprint Corp.'s Sprint E|Solutions and former CTO of pioneer application service provider USinternet- working Inc. "The goal is not to homogenize methodologies, but to provide control points and an end-to-end view, so that system level responsibilities are in someone's hands."

A traditional connectivity service might be composed of multiple network devices: an access device port here, a router, switch and multiplexer port there. In contrast, a managed service multiplies the dimensions -- not only multiple categories of data center and network devices, but multiple services with varied quality guarantees, each one associated with the specific privileges and permissions of specific users.

"The order process leads to identifying all the components in an engineering design plan, and you enter all that in the customer and service profiles," Gastrock says (see chart above). Components are ordered; technicians rack, stack and wire devices; and software is loaded and configured manually and remotely.

But along the way, as different customers want to set up a firewall or hosted message box this way or that, "you can't get all that in a statement of work in advance," he says. "On the front end, you need to recognize the nature of collaborative implementation -- a team with a program manager and a technical lead for a system level view. The band leader oversees all this, and his success is ultimately measured by the promised implementation time, which could be 45 to 90 days for very customer-specific, complex jobs."

Sprint selected a work-flow management middleware from Kintana Inc. to assign and trigger manual and automated tasks. On top of that middleware rides Netcracker Technology Corp. infrastructure asset inventory, Quallaby Corp. service level agreement (SLA) enforcement tools and other OSS applications.

"The implementation OSSs are not islands; they feed historical records to billing, support and other systems like fault management, and a whole bunch of policy devices work in concert to create the contracted user experience," Gastrock says. "If lots of customers want unique policies, it's tough to manage, and right now that's the case: it's hard to manage at the macro level."

Parallel in-house efforts are under way at many service providers including Conxion Corp., Genuity Inc., Telecomputing Inc.'s April 2002 spin-off Apptix ASA and WorldCom subsidiary Digex.

Vendors delivering macro management off-the-shelf solutions include Atreus Systems Inc., Ellacoya Networks Inc., Kintana, Orchestream Holdings plc, Siemens AG's subsidiary Efficient Networks, Step 9 Software Corp. and Telecom Italia-Verizon co-venture Sodalia.

In February, Step 9 introduced its iCustomer SDX service delivery management solution designed to track customers, products and services throughout the service delivery lifecycle. "The challenge is answering how you associate each customer with an array of services and delivery methods, including your own and partner infrastructure, and with the people in your operations," says Step 9 chairman and CEO Michael Rowny. "The goal is any customer can get any product at any time."

Reaching that goal requires control of multiple domains: delivery elements, customer profiles, support processes and business intelligence to diagnose and improve service. But first, Rowny insists, must come control of service definition, including a hierarchy of the myriad products and tasks that underlie each marketed product.

"All of a sudden you may have 10 products you're marketing, and under that hundreds of provisioning products and thousands of provisioning tasks under that, so if you handled each customer as a one-off, the system would fall apart," he says. "Instead you manage your catalog as repeatable processes with defined speed and quality, whether you're provisioning a T1 or a Unix server or a particular flavor of Unix operating system or software application."

In a practical sense, such a catalog of compound service definitions can be made as useful to back-office as to front-office eyes.

Similarly, Atreus Systems' xAuthority Service Fulfillment Solution applies a "define once and automate" Service Framework that marries data on users, services and systems, says product marketing director Brenda Toonders. With pre-built Service Drivers that extend to storage systems, server middleware, unified messaging and other applications from the likes of Microsoft Corp. and iPlanet, as well as to network elements, she says, "service providers can begin to think about creating service bundles through nothing but plug-in modules and the ability to publish them to select, targeted customers."

While most of these systems had entered trials by the winter of 2002 with unnamed service providers, ASP Telecomputing took its own automated provisioning system, called TECOS, and in April spun it out into a new company, Apptix, offering private-label application enablement and support services on a commercial basis to telcos, hosters and system integrators entering the ASP business.

For an upfront fee of $150,000 for integration, followed by an ongoing revenue split , Apptix packages new service creation, along with out-of-the-box support for multiple flavors of hosted Microsoft Exchange unified messaging, Microsoft SharePoint conferencing and collaboration, mobile communications and other applications. 

Share this article: Email, Slashdot, Digg, Del.icio.us, Yahoo!MyWeb, Windows Live Favorites, Furl
Add this article feed to: RSS, My Yahoo, Newsgator, Bloglines