With existing implementation, the hashcode of LispEidRecord
should always be recaluclated when we lookup MapDB.
If we have very large number of entries stored in a HashMap,
the performance becomes very poor.
With this cached hashcode implementation, we can improve
the MapDB lookup performance.
Change-Id: Ie193e0b96b2bdc470e52f6f1de4f341527e0507d
If an application want to create custom Intent installer, it might need
ObjectiveTrackerService.
However, the ObjectiveTrackerService is not public api (in /core/net,
should be /core/api); so application outside the ONOS can't use it.
Move ObjectiveTrackerService and it's dependence to /code/api
Change-Id: Ie2ad8dd9d39ccae1c91f4b57887089e0aec83ede
Javascript's String.split() method is broken.
selArray[1] will not show the full MAC address, instead, it only
shows the first two digits of MAC which is not what we expected.
Change-Id: I4251fd4553d027ae7fb46daf95702bdacc8eeb1e
The meter removal is computed by looking on the MeterStats result
and looking if it is actually missing. MeterStats are retrieved
periodically with the collector. With this patch we ask for the stats
immediately.
Change-Id: I432976d6571d0df4261afdec20a99846dae07f48
Using the MeterStats entries and the actual MeterStore state of the meters notify
just before MeterStore converts the state to ADDED. Remove case is different since
the PENDING_REMOVE state in meter store corresponds to MeterStats meter missing.
Change-Id: I0fe46e1618312b816534c6a7bda7d9e618ca1f7e
- augment UiLinkId with Type field
- preparing for new JSON structure in Topo2Jsonifier
- additional unit tests
Change-Id: Ib8628fc26f72c0369bf8358af905bc6a75e31397
Next-objectives that edited groups are now queued in the FlowObjectiveManager instead of the driver.
During linkup immediately checking for previous portups that should be added to a hash group.
A final retry 30 secs later to catch all ports that should be part of the same hash group.
Change-Id: I7ef450149d685890ca47932b8e559a0c11dc5ab4
Change 12125 in gerrit is causing regression of intent throughput test. The
results of 3, 5 and 7 node cluster regressed about 25%. Single-node results
stay the same (see attachments). NO errors/warnings/exceptions observed in
the log.
The reason is using *compute* method in IntentStore, resolve it by using
other method instead of *compute* method.
Change-Id: I4a062e6c6d5e0898e2e5ac37b2e67a2875d16846
Avoid same IntentData resubmitted by multiple ONOS nodes.
Also, avoid same IntentData/IntentOperationContext modified by different
threds.
Change-Id: Ifb550c41204d6e86e379fd44c2787980920d2b9e
- Enhanced BiLink to allow a UiLinkId as its identifier.
- Corrected aggregation of traffic data for synthetic links in regions.
- Removed TrafficLink copy constructor (not needed).
Change-Id: Ia7d84f28ec69737e7474bfc100262b837effef73
