The Data Transport Problem

Transporting big data over long distances takes too long and costs too much. The current approaches to data transport are expensive, only work over short distances, and produce sub-par throughput.

Typical Solution: Transport the data using traditional delivery methods (portable drives, trucks, planes)

Problem With Solution: Expensive, slow delivery and retrieval, many failure points during transfer, security issues

Typical Solution: Use readily-available TCP/IP links that are cheaper than “dark fiber” to connect locally or globally

Problem With Solution: TCP/IP wasn’t built for long distance data transfer, a few milliseconds of TCP/IP latency (the time to receive a network response) creates dramatic performance degradation, first-generation solutions optimized the wide-area network by changing the data but unstructured data, like medical images, GIS, and high-definition video, cannot be compressed

Typical solution: Use “dark fiber” telecommunications links to improve data transport

Problem With Solution: Very expensive, not available outside of select urban areas, transfers between cities creates hand-offs between carriers, bigger pipes are less efficient than smaller pipes (high latency causes a 1Gb/second link to demonstrate a lower utilization percentage than a 10Mb/second link)