Case Studies : Blue Mug Researches Performance Envelope for Image Converter

Customer: BigCo (we cannot disclose client name)
Industry: Consumer, business and industrial hardware devices
Solution: Performance and feasibility assessment for Image Converter product

BigCo Profile:

BigCo is a large manufacturer of business machines, cameras, optical and other products.

Technical Challenge:

Among other things, BigCo makes printers and cameras. Noting that digital camera silicon storage is expensive (~$0.50/Mb), they came up with the idea of making it simple for customers to off-load images from their cameras to cheap CD-R media (~$0.0005/Mb). Their Image Converter product is a mobile CD-burning unit that connects directly to digital cameras for one-press data transfer. They also considered including a small LCD display and the ability to print directly from the Image Converter device.

To keep their cost-of-goods down, BigCo decided to use a 486-66 processor, a 4xCD burner, and minimal RAM and ROM. Linux was selected as the most attractive OS option because it can be used royalty-free and includes all the necessary software components for connecting to the camera and creating a CD image.

BigCo asked Blue Mug to investigate two issues:

• Could their Linux/486 configuration reliably meet their performance requirements?

• Could this configuration be re-used in other products under consideration, such as an MP3 encoder?

This information would be integral in assessing the product's overall feasibility.

Blue Mug Solution:

Blue Mug assembled a device prototype that would resemble the Image Converter device in all significant regards. We then ran a battery of rigorous performance tests to define the behavior for this device. We concluded that our client could indeed expect good image manipulations and writing of CDs at up to 2x speed, but the CPU simply wasn't fast enough to reliably support 4x writes. Computationally intensive tasks such as MP3 encoding were beyond the abilities of the hardware. We then wrote a report which identifies the usability implications of these findings and proposes a variety of possible solutions.

You can read our report here (MS Word; 105Kb).

Benefits to BigCo:

• Provided quick answer to "Can it be done?"
• Identified bottlenecks
• Researched wide range of solutions and vendors
• Blue Mug's prototype provided a technical blueprint for making the final product