Page 173 - ITU-T Focus Group IMT-2020 Deliverables
P. 173

ITU-T Focus Group IMT-2020 Deliverables                                3


            Action_activateSliceResources:  This  action  causes  a  Slice  to  be  activated  according  to  the  results  of  a
            computation of which resources should be allocated. It will have to configure the fronthaul network, decide
            where the DSP needs to go, allocate the resources both hardware and software for the RAT/DSP create any
            necessary fabric interconnect in the C-RAN, configure any inter C-RAN connectivity, configure any datacenter
            fabric, configure the servers to run the core, pick the software for the core and request it be started, configure
            any IPVPNs and/or L2VPNs, chain together the core elements and of course provide QOS control end to end.
            These would be done by making requests to the CRAN components, DC components and IDC components
            controllers respectively. The error handling here is complex, for example, should the slice come up if all
            resources are not available.
            Action_computeSliceResourceDelta: This would take a running Slice and a new Slice definition based on a
            changed profile and would compute what new resources are needed (or what needs to be released) to make
            the running Slice equal to the new Slice. It does not take any action but simply tries to optimize what the
            changes should be and augments the running Slice with the delta results.

            Action_activateSliceResourceDelta:  This  uses  the  results  of  the  delta  computation  action  to  actually
            allocate/de-allocate resources in the running slice to bring the current slice into agreement with the delta. If
            a RAT / CORE or MEC function is going to be aborted it should be given time to gracefully shut down and
            therefore interfaces would be required to communicate this between the orchestration system and these
            applications.

            Action_shutdownSlice: This would force a slice to be shut down. There are a variety of ways to do this from
            a graceful slow approach to a sudden and abrupt termination and freeing of all resources. The Profile could
            specify shutdown behaviours for various components. The RATs/COREs/MECs functions should be informed
            that they are going down and given time to gracefully shut down prior to being aborted.
            Action_querySlice:  This  event  would  return  the  status  of  the  slice.  It  would  query  the  independent
            components  that  make  up  the  slice  and  would  returned  detailed  status  of  each  individual  component.
            Various wild carded forms of query should be supported to get subsets of the information.
            Action_xyzSlice: A place holder for many more possible actions.

            8.2.3   Action timelines
            This section lists the rough sequence of events that have to occur for a given action at the orchestration layer
            and what the sub controllers would need to do in response to that action.

            Action_connectToCran: This action will attempt to establish a secure tunnel to the controller for the named
            C-RAN.  A  key  exchange  is  done  to  obtain  a  secure  connection  after  which  a  request  is  sent  for  the
            CranCapabilities. These capabilities should contain all the antennas, their locations, properties, name of their
            attachment  points  (what  would  switchable  fronthaul  look  like  outside  the  C-RAN?),  descriptions  of  the
            general  purpose  compute  resources  and  special  purpose  hardware  resources  and  the  fabric  that
            interconnects them together with different RATs available and for each its requirements in terms of compute
            / acceleration components and available spectrum options.
            Action_connectToDran: This action will attempt to establish a secure tunnel to the controller for the named
            D-RAN.  A  key  exchange  is  done  to  obtain  a  secure  connection  after  which  a  request  is  sent  for  the
            DranCapabilities. These capabilities should contain all the antennas, their locations, properties , name of
            their  attachment  points,  descriptions  of  the  general  purpose  compute  resources  and  special  purpose
            hardware resources and the fabric that interconnects them together with different RATs available and for
            each its requirements in terms of compute / acceleration components and available spectrum options. This
            is very similar to a CRAN except much smaller and may have only a few antennas and very limited compute
            resources.
            Action_connectToIDC: This action will attempt to establish a secure tunnel to the controller for the named
            Inter Data Center network (metro etc.). A key exchange is done to obtain a secure connection after which a
            request is sent for the IDCCapabilities. Since the IDC controller may be flexible and able to allocate new
            resources on demand, the Capabilities are not hard upper bounds and may have a flexible representation to
            indicate this. The Capabilities should list the attachment point names such that correlations can be made to
            the attachment point names in the C-RANs and other DC’s or other networks.

                                                                                                         167
   168   169   170   171   172   173   174   175   176   177   178