Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Why do you think it would be <1ms latency saving?

Just my ISPs PoP to Google's closest geo node is ~8ms and that is entirely over fiber and even going through IXPs.



I see it being huge for content caches and CDN's. I see it as more about saving backhaul bandwidth than anything else.


But if you have fiber 100 Gbps is easy and 10 Tbps is possible, at higher cost. This is today, in a year it will be 2 or 4x. This is all a single fiber pair, probably there are dozens.


Storage density growth has long outpaced bandwidth growth, so it makes more sense to have caches everywhere (and Internet protocols that can make use of them), then betting on growing bandwidth.


You might want to minimize the length of fiber you would need (and port costs on the DC side) by connecting from one base station node via another rather than backhauling everything directly to a datacenter somewhere; using edge caching can help reduce the congestion that this will cause.




Consider applying for YC's Winter 2027 batch! Applications are open till November 2.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: