OSM PoC 5 - Placement of Workloads in Distributed Cloud Networks: Difference between revisions
From OSM Public Wiki
(Created page with "= PoC Team Members = Arctos Labs, Telenor, Wind River, Netrounds ''All PoC Team Members are OSM members or participants'' Main Contact: Mats Eriksson (Arctos Labs) [mailto:...") |
No edit summary |
||
(8 intermediate revisions by the same user not shown) | |||
Line 2: | Line 2: | ||
Arctos Labs, Telenor, Wind River, Netrounds | Arctos Labs, Telenor, Wind River, Netrounds | ||
''All PoC Team Members are OSM members or participants'' | ''All PoC Team Members are OSM members or participants'' | ||
* Arctos Labs: Mats Eriksson [mailto:mats.eriksson@ARCTOSLABS.COM mats.eriksson@ARCTOSLABS.COM] - Main Contact | |||
* Telenor: Min Xie [mailto:min.xie@telenor.com min.xie@telenor.com] | |||
* Wind River: Masoud Fatollahy [mailto:Masoud.Fatollahy@windriver.com Masoud.Fatollahy@windriver.com] | |||
* Netrounds: Marcus Friman [mailto:marcus.friman@netrounds.com marcus.friman@netrounds.com] | |||
= Abstract = | = Abstract = | ||
This PoC | This PoC demonstrates a solution that can optimize workload placement over a distributed and interconnected multi-cloud (e.g. edge). The optimization takes technical service constraints into account whilst at the same time finding the most cost optimal placement for each VNF in the service chain, including costs related to their interconnection over a backbone | ||
= PoC Key Takeaways = | = PoC Key Takeaways = | ||
Constraint models complementing NSDs in order to capture service performance requirements: | Constraint models complementing NSDs in order to capture service performance requirements: | ||
* Placement of VNF workloads based on latency requirements | |||
* Placement decisions using real-time latency measurements | |||
* Placement optimization using cost models to predict link and compute costs | |||
* Placement optimization assurance to continuously re-evaluate in case of DC or link failures | |||
= PoC Proposal = | = PoC Proposal = | ||
Line 23: | Line 26: | ||
= OSM Components = | = OSM Components = | ||
* OSM Rel | * OSM Rel FOUR | ||
= Demos = | = Demos = | ||
* EdgeCloudSummit, Paris, 4-6 June 2019 | |||
'''Would you like to see your OSM PoC listed in this wiki? Please follow the instructions in the [[OSM PoC Framework]]''' | |||
'''''Would you like to see your OSM PoC listed in this wiki? Please follow the instructions in the [[OSM PoC Framework]]''''' |
Latest revision as of 09:54, 13 March 2019
PoC Team Members
Arctos Labs, Telenor, Wind River, Netrounds
All PoC Team Members are OSM members or participants
- Arctos Labs: Mats Eriksson mats.eriksson@ARCTOSLABS.COM - Main Contact
- Telenor: Min Xie min.xie@telenor.com
- Wind River: Masoud Fatollahy Masoud.Fatollahy@windriver.com
- Netrounds: Marcus Friman marcus.friman@netrounds.com
Abstract
This PoC demonstrates a solution that can optimize workload placement over a distributed and interconnected multi-cloud (e.g. edge). The optimization takes technical service constraints into account whilst at the same time finding the most cost optimal placement for each VNF in the service chain, including costs related to their interconnection over a backbone
PoC Key Takeaways
Constraint models complementing NSDs in order to capture service performance requirements:
- Placement of VNF workloads based on latency requirements
- Placement decisions using real-time latency measurements
- Placement optimization using cost models to predict link and compute costs
- Placement optimization assurance to continuously re-evaluate in case of DC or link failures
PoC Proposal
PoC Report
OSM Components
- OSM Rel FOUR
Demos
- EdgeCloudSummit, Paris, 4-6 June 2019
Would you like to see your OSM PoC listed in this wiki? Please follow the instructions in the OSM PoC Framework