In preparation of having multiple function classes (e.g., RUST_WASM, WASM_SGX) it makes sense to switch the function execution class at agent level. Doing so allows one node to run multiple runners with potentially different capabilities.
Added a hash map of runners instead of just the default RUST_WASM runner. This can be extended for debug purposes or meaningfully enriched in future deployments. As the AgentRequests are not really consistent with each other (i.e., some provide instanceIds, some functionIds, some also the class), it is also necessary to save the existing functions with their componentId to later find the function class and the respective runner.
In preparation of having multiple function classes (e.g.,
RUST_WASM
,WASM_SGX
) it makes sense to switch the function execution class at agent level. Doing so allows one node to run multiple runners with potentially different capabilities.Added a hash map of runners instead of just the default
RUST_WASM
runner. This can be extended for debug purposes or meaningfully enriched in future deployments. As theAgentRequests
are not really consistent with each other (i.e., some provideinstanceId
s, somefunctionId
s, some also the class), it is also necessary to save the existing functions with theircomponentId
to later find thefunction class
and the respective runner.