Azure vs On-premise Service Fabric -
i have bit of trouble finding differences azure , on-premise service fabric versions. did read somewhere on-premise version not support auto-scaling, easy understand.
however, on-premise version offer type of operational capabilities such resource managers, visual management of cluster, etc.?
the core service fabric platform runtime gets installed on set of virtual or physical machines. once tell machines how find each other, form cluster , provide set of management capabilities includes service fabric explorer ui, rest api, , tcp endpoint powershell. of common whether you're running on azure, on-premises, or in public cloud.
what's different in environments lives outside of machines form cluster. includes:
while service fabric can handle new machines being added , removed cluster, has no knowledge of how process works, external agent needs handle it. in azure, that's virtual machine scale set.
good management of failure , upgrade domains critical ensuring availability , data reliability in service fabric. in azure, clusters automatically spread across fds/uds , maintenance coordinated avoid impact clusters. in other environments, responsibility.
in azure, service fabric cluster 1st class resource can created , managed through azure resource manager , azure portal. outside of azure, must management using cluster configuration json template.
incidentally, there's no confusion since there overloaded terms... can't use azure resource manager (arm) service fabric outside of azure environment. however, service fabric's cluster resource manager part of core runtime , available everywhere.
by default, service fabric logging (on windows) done via etw. however, without component pick events individual machines in cluster , ship them somewhere easy aggregation , inspection, logs aren't useful. in azure, process handled windows azure diagnostics (wad) agent, whereas in other environments responsible setting pipeline.
Comments
Post a Comment