Add workaround for BandwidthChecker issue

[albrow]

Fixes #475.

The current implementation will ban peers if they violate bandwidth limits 4 times over the course of 6 hours. Each violation resets the TTL to 6 hours. If there are no violations over a time span of 6 hours, their count of violations is reset.

This workaround does slightly reduce the resiliency of the Mesh network in the face of spamming attacks, but the impact is minimal and it is much better than having no bandwidth limits in place at all.

[albrow]

I slightly altered the existing BandwidthChecker test. If you run with the -v flag to display logs, you can see that this workaround is working as intended:

go test ./p2p -run BandwidthChecker -v
=== RUN   TestBandwidthChecker
=== PAUSE TestBandwidthChecker
=== CONT  TestBandwidthChecker
time="2019-10-28T13:22:34-07:00" level=warning msg="detected high bandwidth usage" bytesPerSecondIn=787.3617747810655 maxBytesPerSecond=1 numViolations=1 remotePeerID=16Uiu2HAm5BkJwosfgKCJQTz9cT9ewTfakpuM4YVtXLnfKex6TTKf
time="2019-10-28T13:22:34-07:00" level=warning msg="detected high bandwidth usage" bytesPerSecondIn=787.3617747810655 maxBytesPerSecond=1 numViolations=2 remotePeerID=16Uiu2HAm5BkJwosfgKCJQTz9cT9ewTfakpuM4YVtXLnfKex6TTKf
time="2019-10-28T13:22:34-07:00" level=warning msg="detected high bandwidth usage" bytesPerSecondIn=787.3617747810655 maxBytesPerSecond=1 numViolations=3 remotePeerID=16Uiu2HAm5BkJwosfgKCJQTz9cT9ewTfakpuM4YVtXLnfKex6TTKf
time="2019-10-28T13:22:35-07:00" level=warning msg="banning peer due to high bandwidth usage" bytesPerSecondIn=787.3617747810655 maxBytesPerSecond=1 numViolations=4 remotePeerID=16Uiu2HAm5BkJwosfgKCJQTz9cT9ewTfakpuM4YVtXLnfKex6TTKf
--- PASS: TestBandwidthChecker (2.92s)
PASS
ok  	github.com/0xProject/0x-mesh/p2p	2.941s