|
What happens to an enterprise device when it’s retired? Ideally, it’s decommissioned—whether that means wiped of all data and recycled, or outright destroyed.
But with so much hardware to go around, an awful lot of gear can slip into the ether. At DEF CON 32 in Las Vegas this August, Snap Security Engineer Matthew Bryant presented a method that allowed him to identify e-commerce listings for wayward IT assets en masse—including some of Apple’s.
Employing tools like Cloudflare Workers and reverse-engineered APIs, Bryant bypassed rate limiting and scraped 50 million listings from sites like eBay and Xianyu (Chinese shopping platform Taobao’s secondhand market). Bryant hoped they contained clues, like barcodes, indicating anything unusual or sensitive about the devices.
“The challenge is that the secrets we want are probably not outright in the item description,” Bryant told the audience. “Maybe the seller doesn’t even know what they’re selling.”
To extract data from images in bulk, Bryant tested several optical character recognition (OCR) tools. Tesseract, an open-source OCR model, had difficulty with the “very chic, gray-on-the-silver design” used by manufacturers like Apple, Bryant said. Vision, Google Cloud’s OCR API, worked well but was too expensive, he added.
Read the rest here.—TM
|