summaryrefslogtreecommitdiff
path: root/opendc/opendc-compute/src
AgeCommit message (Collapse)Author
2020-04-03Filter on vms updated eventsGeorgios Andreadis
2020-04-01Launch in coroutine scopeGeorgios Andreadis
2020-04-01Optimizes the performance interference model logicGeorgios Andreadis
Credits to Fabian for the idea!
2020-03-31feat: Add reversed policies and random allocation policyFabian Mastenbroek
2020-03-30feat: Add more VM allocation policiesFabian Mastenbroek
2020-03-30bug: Record start and power off events in SC20 experimentsFabian Mastenbroek
2020-03-26Monitor state changes, as wellGeorgios Andreadis
They're now stored in a RLE-fashion.
2020-03-26feat: Implement correlated failures for individual clustersFabian Mastenbroek
2020-03-26feat: Add cluster metadata to bare metal nodesFabian Mastenbroek
2020-03-25bug: Fix race condition in VM provisionerFabian Mastenbroek
This change fixes a race condition in the VM provisioner where VMs were scheduled based on stale information.
2020-03-25feat: Record hypervisor events during experimentFabian Mastenbroek
2020-03-25refactor: Add support for specifying VM nameFabian Mastenbroek
2020-03-25refactor: Migrate to Flow for event listenersFabian Mastenbroek
2020-03-25[ci skip] refactor: Refactor monitors into EventFlowFabian Mastenbroek
2020-03-25refactor: Rework monitor interfacesFabian Mastenbroek
2020-03-25[ci skip] feat: Add support for correlated failuresFabian Mastenbroek
2020-03-25feat: Use Weilbull distribution for failuresFabian Mastenbroek
2020-03-25feat: Add failures to SC20 experimentFabian Mastenbroek
2020-03-25feat: Track VM failures in SC20 experimentFabian Mastenbroek
2020-03-25feat: Add support for failing hypervisors in provisionerFabian Mastenbroek
2020-03-25feat: Make FaultInjector more genericFabian Mastenbroek
2020-03-25[ci skip] feat: Prototype design for FailureInjectorFabian Mastenbroek
2020-03-25feat: Propagate machine failure to VMFabian Mastenbroek
2020-03-25feat: Add infrastructure for failuresFabian Mastenbroek
2020-03-25feat: Revamp bare-metal state managementFabian Mastenbroek
2020-03-25feat: Terminate server instance on power offFabian Mastenbroek
2020-03-16Extend performance interference model to support minimum server loadGeorgios Andreadis
Vincent pointed out that performance interference is only triggered at a certain total server load level. This addition supports that.
2020-03-13refactor: Change burst to MFLOpsFabian Mastenbroek
2020-03-13feat: Divide workload equally over pCPUsFabian Mastenbroek
2020-03-12feat: Add VM scheduler for scheduling over all coresFabian Mastenbroek
2020-03-11Cache burst sumGeorgios Andreadis
2020-03-11Fix remainder/burst granted issueGeorgios Andreadis
2020-03-11Merge hypervisor monitor callsGeorgios Andreadis
2020-03-11feat: Use linear power model for SC20 experimentsFabian Mastenbroek
2020-03-11refactor: Extract Signal Flow into odcsimFabian Mastenbroek
2020-03-11feat: Measure host CPU usage and power consumptionFabian Mastenbroek
2020-03-11bug: Propagate remaining burst correctlyFabian Mastenbroek
2020-03-11feat: Provide access to BareMetalDriver in ServerFabian Mastenbroek
2020-03-10feat: Implement basic power modelFabian Mastenbroek
2020-03-10feat: Represent machine load as flowFabian Mastenbroek
2020-03-09feat: Define key for performance interference model of imageFabian Mastenbroek
2020-03-04Fix rebaseGeorgios Andreadis
2020-03-04Make onExit suspendingGeorgios Andreadis
2020-03-04Give vCPUs access to their hosted VMGeorgios Andreadis
This should make performance interference based onn the colocated VMs possible.
2020-03-03refactor: Rename maxUsage to limitFabian Mastenbroek
This change renames the `maxUsage` parameter to `limit` in order to align terminology with other products/projects such as VMWare vSphere.
2020-03-03feat: Re-add support for hypervisor monitorFabian Mastenbroek
2020-02-29refactor: Store remainder in burst arrayFabian Mastenbroek
This change modifies the API of ServerContext to store the remainder burst into the input burst array instead of returning a new array.
2020-02-29refactor: Report CPU usage per serverFabian Mastenbroek
This change refactors the codebase so that the CPU usage of the server is only reported per server, instead of per CPU reducing the total amount of messages needed and additionally simplifying synchronization of various computations.
2020-02-28refactor: Create distinction between CPU node and coreFabian Mastenbroek
This change updates the terminology in the `opendc-compute` module to make a distinction between CPU node and CPU core, where we primarly work with CPU cores. However, if needed, we also provide information for the different CPU nodes.
2020-02-28Fix NPEGeorgios Andreadis