v1.1.0--fix the problem of too long route synchronization time, and update the advertise-service-cluster-ip-range function.

--The function of advertise-service-cluster-ip-range has been updated.
--bgp-graceful-restart-ipv6 function is added to solve the problem of repeated packet transmission in the IPv4 scenario.
--Also dep is no longer used, we will be using go modules going forward for dependency management.
--Go 1.13 is version used for building kube-router and gobgp binaries.
--Starting from this release kube-router by default tries to auto-configure MTU for both pod interfaces and kube-bridge bridge interface.
--Fix the problem of too long route synchronization time.
--k8s.io/client-go, k8s.io/apimachinery, k8s.io/api are updated to v0.18.8.
--gobgp is updated latest version 2.19.
--Fixed some bugs.

v1.0.0--The bgp-holdtime function parameter of setting holdtime is added to adjust the holdtime of BGP negotiation with the connected network devices.

In some scenarios, the holdtime of BGP needs to be adjusted.
(1) If you want the BGP neighbor route to disappear as soon as possible after the kube-router goes down, you should set the holdtime to be smaller.
(2) In some scenarios, after the kube-router goes down, the pod network is normal and can still communicate with the outside world. You want BGP neighbor routing to last longer, so in this case you can set the holdtime to be larger.

v0.3.2--This version optimizes the cluster subnet announcement and supports large-scale networks

This version optimizes the cluster subnet announcement and supports large-scale networks.We have done the following based on kube-router:
--Support large-scale kubernetes network.
--Advertise cluster subnet to BGP peer.We added the "advertise-cluster-subnet" parameter to announce the specified service subnet route.
--Optimize BGP route update.