Cover Story: BCP Strategies
Taking a call on continuity

The Global eCMS call center has 1000 seats and caters to eight clients worldwide. A minute of downtime or interruption in its services will result in financial losses and irreparable damage to its reputation. The company has deployed redundancy, contingency, and recovery solutions at various levels to ensure as much uninterrupted service as possible. by Soutiman Das Gupta

The Global eCMS call center in Mahape (Navi Mumbai) currently caters to eight clients and has a capacity of 1000 seats. A large team of call center operators ensure around the clock service at the facility. Any downtime or disruption of services will lead to direct loss of revenue and loss of reputation. In such a scenario BC planning and solutions are absolutely essential.

In the area of BC solutions, Savio Furtado, Senior Vice President, Technology, feels that the organization has a very high MTBF (Mean Time Between Failures) of 10 years on all critical equipment. MTBF is a measure of the amount of time a product is likely to function without failures. The higher the MTBF, the more reliable the product. He also feels that the company has deployed very reliable redundancy, continuity, and disaster recovery plans.

Areas of concern
Global eCMS's areas of concern are:

• Voice systems
• Call Management systems
• CRM/Dialer
• The call center LAN
• The Web servers
• The last mile

Each area has been carefully looked into and continuity solutions have been implemented at most levels.

Voice systems
A call center's most critical business need is undoubtedly serviced by the voice systems setup. So it was only natural that Savio considers this the highest area of concern.

The voice systems have been further classified into technology areas to help maintain better continuity planning. The areas are:

• PBX - Private Branch Exchange
• MUX - Multiplexing systems
• IPLC - International Private Leased Circuit
• Co-location equipment
• PSTN - Public Switched Telephone Network

For PBX-related functions, the company plans to upgrade the SCCS (Symposium Call Center Server) software from Nortel, to version 4.2 by December 31, this year. The power supply module and the CP is redundant in nature. A new PBX will be deployed for the company's USA business by the end of March 2003. The existing PBX will serve the UK business. And both the PBXs will act as a fail-over to each other.

A second site essentially for disaster and fail-over will be set up in Pune, and is expected to be completed soon. Pune was chosen because it is in a different seismic zone. Spare stock of equipment and spares are kept at all locations. The voice multiplexing equipment has redundant systems for its OS, power supply, MVP modules, and protocols.

Global eCMS already uses IPLC lines to its co-locations in UK and USA from carriers like AT&T, MCI, BT, and C&W Telecom. The lines have adequate bandwidth to support communication between the locations. There are multiple lines that send data through various routes (Atlantic and Pacific). Lines from SEMEWE3 and FLAG Telecom are used for backup.

Evaluation for backup co-location facilities at USA is in the process. Backup co-location facilities in UK will be set up in 2003. The call center uses multiple service providers for PSTN links. There is a mix and match of E1, T1, E3, and T3 links for redundancy and fail-over.

Call Management System
Global eCMS uses the ACD functionality of the Nortel 81C in conjunction with the SCCS to cater to its call management requirements. The 81C is an enterprise class equipment and has inbuilt redundancy in terms of power supply and the CP.

CRM/Dialer
The call center uses products from Cincom/Davox (Concerto) for its CRM/Dialer needs. It enables the center to perform preview dialing and predictive dialing. Davox Ensemble provides blending capabilities. MS SQL and Sybase are used as the RDBMS and runs on redundant NT and Sun servers. The Dialers are currently located in India and a new Dialer is being installed in USA.

The LAN
The facility's LAN has a number of devices like core switches and floor switches. The backbone is optical fiber. There are two switches at the core and the first three floors have two switches for redundancy. The third floor and basement have single switches.

Web servers
Web servers are set up in a clustered configuration for fail-over and backup. The links to the clusters are regularly tested for performance issues. The servers operate in a secure mode using the SSL 128-bit encryption.

Last Mile
Fiber links from MTNL connect the last mile to the call center through redundant exchanges at Turbhe (Navi Mumbai), Millennium Business Park and Prabhadevi (Mumbai). The links are in a mesh and in case one link is down, the data is routed through another link. Currently MTNL switches between fiber manually. By October MTNL plans to introduce automatic switching that can be managed online.

Soutiman Das Gupta can be reached at soutimand@networkmagazineindia.com