Skip to content

APIGW vs. v5c GW

mmatsa edited this page Feb 20, 2019 · 3 revisions

APIC v2018 provides fundamental changes over APIC v5, including a new Gateway APIGW which provides fundamental changes over API v5's Gateway. Note that APIC v2018 actually provides two Gateway options for APIs, the v5-compatible DataPower Gateway aka v5c and the DataPower API Gateway aka APIGW.

Overall, there are a few high-level concepts that are useful for understanding the shift from APIC v5 to the new API Gateway.

#1 : Native DataPower Gateway

The v5c GW is built on top of XSLT and GWS which are running on a DP MPGW:

image

The APIGW is a native DataPower service-level object, alongside MPGW, so API policies are natively implemented and run directly:

image

This means that the mapping from APIC concepts to Gateway implementations is direct, "What you see is what you get", which implies a much easier debugging and maintenance experience. It also yields algorithmicly faster performance.

#2 : Horseshoe Flow

The v5c GW is built with a request/response flow:

image

The APIGW is designed with a "horseshoe" flow, thus allowing first-class patterns with no backend at all, one backend, or many invoked backends all gathering data for the response to the client:

image

#3 : Gateway Cluster managed on Gateway-side

In APIC v5, the Management node knows about every Gateway in the Gateway cluster and manages the set:

image

In APIC v2018 (for both Gateway types), the Management node knows only about the Gateway Cluster as a whole, and the cluster is in control of members of the cluster itself. This means that the administrator in charge of the Gateways can take individual Gateways in or out of rotation, scale-up/scale-down, etc., without needing to touch the Management node.

image

Links

For a comparison of the v5c GW and the APIGW, see https://www.ibm.com/support/knowledgecenter/en/SSMNED_2018/com.ibm.apic.overview.doc/rapic_gateway_types.html

Clone this wiki locally