Will widespread edge computing lead to widespread edge interconnection of the sort that presently happens in big data centers? Yes, but “where” interconnection happens will depend on what sort of “edge” we are talking about. Some larger facilities that are metro level will certainly support interconnection.
Lumen Technologies, for example, believes it can cover about 98 percent of edge computing demand with five millisecond or less latency, from a relative handful of regional locations. Others networks of small edge data centers, spaced at about 20 kilometer distances, also will work.
IDC has estimated that 75 percent of enterprise latency needs are satisfied with latency of less than five milliseconds. About 15 percent of use cases can tolerate latency of five to 10 milliseconds.
No comments:
Post a Comment