Printer Friendly

Fiserv banks on dial-up rather than leased lines: bridges point-to-point lines into virtual multipoint circuits.


Many banks, credit unions and savings institutions rely on service bureaus for their data processing needs. They also rely on these service bureaus to have comprehensive disaster recovery plans to assure the continued operation of the institution's data processing operations during times of natural disaster or system breadown.

The greater transaction speed and complexity brought on by the computerization of financial institutions means the loss of data, or the complete breakdown of the transactional system, is at a greater risk from disaster.

Fiserv, Inc., with headquarters in Milwaukee, provides specialized financial data processing services and related information management systems to financial institutions nationwide. Through its 20 data centers across the U.S., Fiserv services over 19 million accounts for more than 800 financial institutions. It's the nation's largest DP services to savings institutions, and ranks among the leading independent servicers in both the commercial bank and credit union industries nationwide.

Each data center in the Fiserv system is responsible for its own bottom line, and though an ever increasing national standardization is providing significant economies to the individual operations, each center continues to play an active role in its own hardware and software decisions. As centers with common hardware platforms become increasingly standardized, more and more decisions are made in concert, such as the contengency plans created for the Fiserv centers with Unisys mainframes that are processing banks and savings and loans.

Heavy Respon ibility

"There is a definite responsibility and a cautious self awareness that comes with our position in the financial services industry," says Mike Rigney, vice president of operations at the Fiser center in Tampa, Fla. "Because of our national presence and market position, we share heavily in the responsibility for being prepared to contain a disruption in service before it results in a long-term impact on our clients or their customers."

Recognition of the financial industry's increasing dependency on data automation has prompted federal financial examiners to renew emphasis on disaster preparedness in the form of tough new compliance reqirements for financial institutions and servicers. Even before that, Fiserv went through the process of designing and implementing plans for computer hardware and facility recovery, as well as establishing an organizational structure to deal with disaster recovery.

"That was the easy part," Rigney says. "The most difficult job was coming up with an effective means of providing data communications access between end-users and the recovery site following disruption." In the service bureau environment, a centrally located data center may service a region of 10 or more states using a multidrop data line network (a four-wire leased telephone line that provides terminations at multiple geographically separated locations) to communicate on-line transactions with client locations.

Nixed Overkill Option

Fiserv looked at AT&T's T1 Accunet Reserve, but Rigney describes that application as a "Big Bang" approach. The Accunet Reserve application consisted of a network of T1 multiplexers acting as hubbing points for existing multipoint analog and digital tail circuits to the end user.

The primary advantage of the Accunet method was its global scope and its ability to reroute every user to a disaster recovery center with not involvement from the user. For the purposes of Fiserv, this method had a couple of key disadvantages:

* Time. The nodal architecture would have required that Fiserv negotiate arrangements with AT&T or the regional operating companies to allow Fiser-owned hardware to be resident on telephone company property, or install alternate master drops on each data line at a local contingency hub site linked to Accunet Reserve. Even then, actual use is subject to availability.

* Cost. T1 subscription and alternate routing expenses are high, and require an "all or nothing" approach to supporting user locations.

As government regulators continued to press teh industry for stricter compliance standards, Fiserv explored options to help client institutions meet the need for contingency planning. "We were driven to look for a more viable, cost-effective solution," Rigney says. "For our Unisys centers that solution was dial back-up."

Fiserv had previously looked at dial back-up as a solution, but it was concerned about going from a multidrop leased-line environment to dial back-up. Converting each multidrop location to a dial point-to-point location would have multiplied the number of front-end processor ports that are necessary at the host site.

Bridging Discovery

In addressing the problem Fiserv consulted Networks, Inc., a data communications system integrator located in Miami Lakes, Fla., to research the marketplace.

Bob Toombs, president of Networks, found a PBX device capable, when used with the proper modem, of allowing numerous two-wire single call point-to-point dial lines to be re-merged (bridged) into a single output line to emulate multidrop environments.

"Adding four-wire, two call back-up to the leased line modems was considered and rejected becasue the ancillary dial back-up hardware cost was not materially lower than the supplementary single call V.32 modems," says Toombs. "The additional recurring cost of two dial lines per site versus one sealed the decision in favor of V.32 modems."

For Fiserv, the advantages include:

* Cost. Because the plan allowed for selective implementation in key user locations, the cost was less than one-third of he T1 subscription cost over three years.

* Service restoral. Because the plan places an additional modem at the user site with dial capability, it is possible to restore service to a user suffering from a failed leased line for leased line modems. T1 would have allowed for recovery from a back-up center, but not restoral to the local center.

Tooms field-tested the dial back-up with Fiserv personnel and client terminals in a network of synchronous, asynchronous and isochronous polled terminals from about 25 different vendors.

Big Challenges

Tooms realized he had to find a system which could physically handle many lines to one-line bridging and accommodate the many sunch and asynch speeds in the systems parallel environment. "We were able to locate a bridging system which would allow combining point-to-point lines into virtual multi-point circuits and find a modem that not only would pass the necessary control signals to the bridging device, but also allow that to be done in a single dial-up call environment," he says.

The next challenge was identifying a modem that was capable of running synchronous, asynchronous and isohronous at any speed from 1200 to 9600 b/s. What was needed was a modem that could be used in any client application in a dial-up environment, and that could be rebridged at the recovery site to fit back into a practical number of front-end processor ports and mainframes.

Toombs chose the NEC 9631 V.32 modem because it can operate in all three modes supported by Fiserv, and at five speeds. The 9631 supported the simulated switched carrier function necessary to emulate a drop on a multipoint circuit.
COPYRIGHT 1990 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Milwaukee, WI, financial data processing firm Fiserve Inc uses dial back-up for reliability
Publication:Communications News
Date:Apr 1, 1990
Previous Article:Schwab ensures voice reliability; customer satisfaction linked to call wait time.
Next Article:Shorten the road to recovery: scheme features computerized blueprint as key to fast on-site restoration.

Related Articles
DP bureau backs S&L customers.
Bike manufacturer rides out problems with analog lines.
InView selects VideoServer.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters