Skip to end of metadata
Go to start of metadata

Blueprint:  Multi-Server Integrated Cloud native NFV/App stack (Short term: MICN)

Project Technical Lead:  Kuralamdhan Ramakrishnan, Elected 6/24/2019

Link to process-SC review record Graduation reviews

Attributes info 

Case Attributes

Description

Informational

Type

Blueprint


Blueprint Family - Proposed Name

Integrated Cloud Native NFV stack


Use Case

SDWAN,  Customer Edge,  Edge Clouds – deploy VNFs/CNFs and applications as micro-services


Blueprint proposed Name

Multi-server Cloud Native stack


Initial POD Cost (capex)

50K minimum


Scale & Type

Minimum of 4 Xeon Servers + 1 Xeon server as genesis


Applications

SDWAN,  ML/DL Analytics, EdgeXFoundry and 360 degree Video streaming


Power Restrictions



Orchestration

Infratructure Orchestration

Bare Metal Provisioning :  ironic with Metal3 controlled by Cluster API

Kubernetes provisioning :  KuD.

Centralized controller:  With Cluster-API

Docker for containers and Virtlet for VMs

Service Orchestration : ONAP (Only the components that require VNF/CNF workload deployments via K8S)

MEC framework: OpenNESS

Site orchestrator :  Kubernetes upstream

Traffic Orchestration within a cluster: ISTIO

Traffic orchestration with external entities : ISTIO-ingress and ISTIO-egress with MCDeployment

Knative for function orchestration

Expose HW accelerators

Storage Orchestration: Ceph with Rook

Futrue: AF-XDP for packet processing based containers, OpenShift for site orchestrator, Kubevirt for VMs.

SDN

OVN, SRIOV, Flannel


Workload Type

Containers, VMs and functions


Additional Details

Future:

eBPF based CNI (such as PolyCube)


System level architect/lead:  Srinivasa Addepalli

Contributors:

CommitterCommitter CompanyCommitter Contact InfoCommitter BioCommitter PictureSelf Nominate for PTL (Y/N)
Kuralamdhan RamakrishnanIntelkuralamudhan.ramakrishnan@intel.com

• Kuralamudhan Ramakrishnan is working as a Senior Software Engineer in Intel, Oregon has specialized in the field of Networking, storage and data management and worked in Telco industries.
• He created various open source projects such as Multi Network plugin in CNI called "Multus" CNI to support Multi-Homed pod in Kubernetes and "SRIOV-DPDK CNI" and "Userspace CNI" working in Kubernetes Networking in Intel, worked on Kubernetes compute on power management and worked in Kubernetes network SIG for Network Plumbing Working Group to advocate for Telco, comms use case in K8s and formation of Multi-network de-facto standard spec in K8s.

• Kural's LinkedIn - link

Yes
Ochulor, Enyinna Intelenyinna.ochulor@intel.com

Enyinna Ochulor is a Cloud Software Engineer at Intel. Since joining Intel, Enyinna has contributed to the OpenStack Neutron (DVR), and ONAP vCPE projects. LinkedIn




Ukponmwan, Itohan Intelitohan.ukponmwan@intel.com

Itohan Ukponmwan is a Cloud Software Engineer at Intel. She has contributed to various projects in ONAP. She worked on the vCPE project in ONAP amsterdam release and contributed to making it possible to Deploy Intel's FlexRAN VNF with ONAP. She has also done a lot of work in enabling HPA features in ONAP for which Intel is a major contributor. Itohan has also contributed and worked on fixing bugs in the Opendaylight controller project.

LinkedIn



Le, Huifeng Intelhuifeng.le@intel.comHuifeng Le is a Cloud Software Engineer at Intel. He has contributed to cloud/network related open source projects such as Openstack Neutron, StarlingX, OpenNESS etc. and openstack service containerization.

Li, ChengIntelcheng1.li@intel.comCheng Li is a Cloud Software Engineer at Intel, has 3 years experiences on OpenStack, focus on neutron/cinder project. Also work on openstack automation deployment, i.e. ansible, docker/kubernetes, openstack-helm, Airship.

Yao, LeIntelle.yao@intel.comLe Yao is a Cloud Software Engineer at Intel. Focus on neutron and kubernetes, virtlet and kubevirt.

Ramki Krishnan

VMwareramkik@vmware.com


Constantine Polychronopoluous VMwarecpolychronop@vmware.com


Amar KapadiaAarna Networksakapadia@aarnanetworks.comLinkedIn

Sriram RupanaguntaAarna Networkssrupanagunta@aarnanetworks.comLinkedIn

Prem Sankar GLumina Networkspremsankar@gmail.comLinkedIN

Ravi ChunduruVerizonravi.chunduru@verizon.comhttps://www.linkedin.com/in/ravivsn/


Chen, TingjieInteltingjie.chen@intel.comTingjie Chen is a Cloud Software Engineer at Intel, he has contributed to edge and storage related open source project such as Kubernetes RSD CSI, StarlingX etc, and also working on Ceph containerization in StarlingX.

Ramamani YeleswarapuIntelramamani.yeleswarapu@intel.com


Haibin HuangIntelhaibin.huang@intel.comHaibin is  Cloud Software Engineer working out of Intel ShangHai. He has contributed to ONAP projects.

Rajamohan RajIntelrajamohan.raj@intel.comRaj is a Cloud Software Engineer at intel. He is currently working on the Distributed Analytics as a Service project.

Igor Duarte CardosoInteligor.duarte.cardoso@intel.comCloud, networks, telecommunications. OpenStack, OSM, Kubernetes, Akraino.

Files

  File Modified
Microsoft Powerpoint Presentation Akraino-Cloud-Native-Edge-Stack-external-v2.pptx Jul 25, 2019 by Srinivasa Addepalli
Microsoft Powerpoint Presentation Akraino-Cloud-Native-Edge-Stack-external-v3.pptx Aug 19, 2019 by Srinivasa Addepalli
Microsoft Powerpoint Presentation Akraino-Intergrated-Cloud-Native-meetup-external-v4.pptx Dec 03, 2019 by Kuralamudhan Ramakrishnan
Microsoft Powerpoint Presentation Akraino-Edge-Service-matching.pptx Apr 21, 2020 by Srinivasa Addepalli



  • No labels

6 Comments

  1. ramki krishnan ,  @Prem, @Amar Kapadia and @Ravi Chunduru,


    Please add your information to above table.


    Srini

  2. Hi Srinivasa Addepalli

    Do we need to have hardware accelerators in the servers for this set-up. Will this work on Dell servers without acceleration?

  3. Yes. It should work even without any accelerators. Only requirement is that servers support IPMI for now.  We plan to support Redfish also eventually.   Using IPMI, it controls BIOS via BMC to setup PXEBOOT, reboot and other operations.  ICN infrastructure orchestration will take care of installing Ubuntu 18.04 LTS version and followed by Kubernetes via KuD (wihch internally uses Kubespray).   It also installs ONAP4K8s.  Then you can deploy applications using ONAP4K8s on multiple Edge K8s clusters. Let us know if you hit any issues.  Make sure to copy Kuralamudhan Ramakrishnan

  4. Thanks Srinivasa Addepalli Kuralamudhan Ramakrishnan

    I see that ICN covers pretty good use cases like SD-WAN, Distributed analytics etc.

    Is ICN also suitable for RAN use cases like RIC or CU sitting on edge. Because these applications may very require high throughput as they are closest virtualized applications to  the users. 

    1. Yes, it is also suitable for RAN use cases. I am working on FlexRAN with N3000 acceleration.

    2. Ankit Arora

      Yes. ICN is for RAN too as RAN is becoming disaggregated.

      We have written one article on ICN applicability for 5G RAN.

      You can find it here:  https://www.linkedin.com/pulse/lfeakrainoicn-role-5g-ran-srinivasa-addepalli

      Thanks
      Srini