Abstract

Today’s Internet naming system is founded principally on the notion that a human needs to communicate with a remote, well-known server. The Internet’s Domain Name System (DNS) works extremely well in that context; for example: a user Alice can simply type www. google. com into their web browser, or bob@ comcast. net into their E-Mail software and “magic happens” in her eyes as the web page is displayed or the E-Mail is delivered. But if you ask Alice to share a file with Bob, or print to her printer when she’s not at home those tasks also require a centralized server infrastructure when they otherwise shouldn’t. The root of this problem is grounded in how Internet naming is constructed today. The Internet’s current naming system (the DNS) falls short of meeting the needs of peer-topeer and peer-to-device based connections. Alice has no way to simple “send this file to Bob’s cell phone”, primarily because Bob’s phone doesn’t have a stable name. Similarly,“send this document to the printer in the den” is equally as impossible, especially when Alice is not within the same network where mDNS [2] no longer works. The result is that Alice must fall back to sharing files via cloud-based services like DropBox [3], printing via Google Cloud Print [5] and even configuring the thermostat across the room using external middle-ware services like EcoBee [4].(Note that technologies exist to send files to near-by devices, such as over bluetooth, the Wifi-Alliance’s Wifi-Direct [8] or Apple’s proprietary AirDrop [1]. However, none of these are cross-platform.) None of these scenarios actually require a round-trip communication through the cloud, but they all do go …

Date

2016

Authors

Wes Hardaker