Determine acceptable bounds on bulk fetch slowdown

We're designing the ChainSync Jumping optimization with the goal of offsetting the cost of maintaining 10s of additional upstream peers compared to today's node.

While we're still optimistic: it remains a possibility that---especially for the first MVP---CSJ might not fully offset that cost. Hence, the question that this Issue is intended to answer:

> Assuming the first fully-featured P2P/Genesis node will be slower at bulk fetch than the preceding node release, how much slower would be acceptable?

Bulk fetch (of the whole chain) already takes several hours. Would eg 30% more be acceptable? What are the tiers of performance/User Experience that we're trying to navigate? Etc.