Networking giant Cisco is spending $655m to fix defective memory chips that it incorporated into a range of its enterprise networking products between 2005 and 2010.
The memory chips, which are essential to the functioning of the products, sourced from a single supplier.
"These components are widely used across the industry and are included in a number of Cisco products. They are known to slowly degrade over time, and in some cases, have caused products to fail after a power cycle event. Failure rates due to this issue have been and are expected to be low," Cisco revealed in its most recent quarterly results.
The supplier of the defective part hasn't been named.
"However, recently Cisco has seen a handful of its customers experience a growing number of failures in their networks as a result of this component problem. Accordingly, although the majority of these products are beyond Cisco's warranty terms, Cisco is proactively working with customers on mitigation, resulting in a charge to product cost of sales during the second quarter of fiscal 2014."
Cisco is keen to assert that it is not the only company affected by the degrading memory problem and that a number of other suppliers are also affected.
The company says that although many of the products affected are out of warranty, it is nevertheless replacing the defective parts.
"The majority of Cisco products using these components are experiencing field failure rates below expected levels. Recently, however, a handful of our customers have experienced a higher number of failures, leading us to change our approach to managing this issue," wrote Curtis Hill, vice president of technical services at Cisco, in a blog posting.
The components are commonly used as the main memory of the processor for the operating system and include memory modules and discrete on-board memory (individual components soldered on a line card or similar printed circuit board assembly), says the company.
Typically, the memory components are more than two or three years old and suffer a functional failure when the devices are switched off and on. The memory failure is attributed to a degradation mechanism in a specific memory circuit design.
"Customers should contact the Cisco Technical Assistance Center (TAC) if they experience a failure in a Cisco product. Cisco is working on making tools available to determine if potentially affected products are in a customer's network," advises the company on its website.
Sometimes, the power of the mainframe is the most cost effective answer. Computing's Peter Gothard puts Computing's readers' questions on the future of the mainframe to IBM's Z13 expert Steven Dickens.
This Dummies white paper will help you better understand business process management (BPM)