Proposed approach to reduce delays.
Proposed approach to reduce delays.
Beyond load balancing, it seems unusual or impossible for two ISPs to reduce latency by half. Failover boosts connection reliability while load balancing can increase data size, but combining both services could theoretically help. Initially, changes might not follow physical laws, but over time both providers could share information. Your idea involves one provider initiating a ping first, then the other follows, exchanging non-own pings instead of direct measurements.
The only way I understood this would be if both ISPs were closely linked together and your data was prioritized as VIP across both networks. It’s technically feasible, but in reality and with your resources it wouldn’t work.
Latency tends to match what you expect based on the underlying physics. The only exception is if one ISP offers a faster path than another, making choosing it technically advantageous. The main issue is that you can never be certain which ISP provides the best connection for any particular destination. If one consistently outperforms another for a specific location, you might implement a router policy to always route through that ISP—such as sending traffic through a US VPN for certain geo-restricted sites. However, in practice, this requires significant effort and routing changes can occur unexpectedly, meaning the fastest ISP isn’t always guaranteed.