Gateway |
support multiple instances of gateway class |
|
doing |
|
support for namespace |
|
doing |
|
support default gateway class |
|
doing |
|
all data stream could load-balance to all gateway node |
|
doing |
|
when a gateway node breakdown , all data stream could to to healthy gateway node |
|
doing |
|
could specify the node interface for tunnel, by hand, or auto select a reasonable one |
|
doing |
Tunnel protocol |
vxlan |
|
doing |
|
geneve |
|
|
Encryption |
ipsec |
|
|
|
wireGuard |
|
|
Destination CIDR |
could auto distinguish internal CIDR (calico, flannel etc, or by hand) and outside CIDR |
|
doing |
|
could specify the outside CIDR by hands |
|
doing |
Data protocol |
tcp |
|
doing |
|
udp |
|
doing |
|
websocket |
|
|
|
sctp |
|
|
|
multicast |
|
|
Policy |
support priority |
|
doing |
|
support cluster scope policy |
|
doing |
|
support namespace scope policy |
|
doing |
Support CNI |
calico |
|
doing |
|
flannel |
|
doing |
|
weave |
|
doing |
|
macvlan+spiderpool |
|
doing |
IP Stack |
ipv4-only |
|
doing |
|
ipv6-only |
|
doing |
|
dual stack |
|
doing |
Source IP |
support EIP for application |
|
doing |
|
support EIP for namespace |
|
doing |
|
use node ip |
|
doing |
Datapath |
iptables with low and high version |
|
doing |
|
ebpf |
|
|
Performance |
big cluster, with lots of gateway nodes |
|
|
|
big cluster, with lots of nodes |
|
|
|
big cluster, with lots of pods |
|
|
|
when gateway node down, the whole cluster could change to healthy gateway node within 2s |
|
|
|
after apply or modify lots of policy, it could quick take effect in a big cluster |
|
|
|
after apply or modify gateway node, it could quick take effect in a big cluster |
|
|
|
forward throughput of each gateway node |
|
|
|
CPU and memory usage under pressure |
|
|
HA |
all component pods could recovery quickly and serve after breakdown |
|
|
|
all pods could run for one week without failure |
|
|
Insight |
metrics |
|
|
|
log |
|
|
Doc |
design, usage, debug docs |
|
|
Architecture |
amd and arm |
|
|