WA infrastructure dept close to e-mail disaster

Western Australia's Department of Planning and Infrastructure (DPI) has revealed its Microsoft Exchange-based e-mail system is suffering frequent outages that are creating a risk of embarrassing public data loss.

"The existing messaging environment suffers from frequent outages (more than 24 hours per annum) due to a combination of hardware and software failures, e-mail corruption and human error," the department wrote in tender documents recently issued with the express purpose of calling for external assistance.

The department's role is to deliver land and transport infrastructure and services to the state. It had close to 1700 employees, as at 30 June last year.

DPI stated its e-mail system had high management overhead and licensing requirements, and "no form of failure protection of disaster recovery". The environment consists of four separate Microsoft Exchange 2003 servers, running on Windows Server 2003 Enterprise Service Pack 1 on Dell hardware and a Hitachi Data Systems storage area network.

The servers are dispersed between four different locations, but all link into DPI's Murray St, Perth offices.

"Current failure rates are increasing, which increases the risk of prolonged outages of the Contract Authority's messaging system," the documents stated. "Not only is the risk of prolonged outages increasing, the risk of data loss through inadequate management, monitoring and backup is increasing dramatically."

DPI noted that although data loss and system outages may not have any current penalties to its operations, there was a "high" risk of potential damage to the reputation of its Contracts Authority division.

"The Contract Authority is responsible for the security and management of many public records, and as such risks serious integrity loss in the event of a disaster or site failure," the tender documents said.

Neither the department nor Microsoft commented on the issue by press time.

The solution
To avoid problems in future, DPI is seeking to implement a "high-availability" messaging replacement for its cantankerous system. The four servers will be cut down to just two -- one primary and one secondary Exchange system. The two servers will be located in different locations, with backup data links.

Both servers will replicate data between each other, and fail-over to each other or to two backup servers in the event of major problems.

This new system would provide the Contract Authority with "a more resilient messaging system, with lower ongoing costs than the one currently in place," the documents stated.