Edge and cloud and MultiCluster Orchestrator Overview Roadmap

Edge (and cloud) and Multi-Cluster Orchestrator Overview & Roadmap Contact: Srinivasa. r. addepalli@intel. com EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO 1

Traditional Cloud Native frameworks For Enterprise applications Apps K 8 s Linux VMs Private Cloud K 8 s cluster FW - K 8 s clusters are used for normal applications - Network and security functions are deployed outside of K 8 s clusters as physical appliances or virtual appliances, typically supplied by providers. Apps Linux SLB Traditional: - Number of K 8 s clusters are small - K 8 s Cluster installation/upgrades are mostly done independently in each location. - Deployment of applications on K 8 s clusters is also done independently. Public Cloud K 8 s Cluster IDS FW Provider supplied SLB IDS Provider supplied Internet EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Today K 8 s Clusters are not meant for Network functions. Need for comprehensive converged Edge platforms.

Trend : Geo Distributed Computing trend with Edge-computing Public/ Private Cloud m. S 4 External System m. S 3 Cloud Platform Public/ Private Cloud m. S 4 m. S 3 m. S 2 m. S 1 Cloud Platform WAN Drivers • • Latency/ Physics Edge 1 Bandwidth/ Economics Context/ Proximity Privacy/ Legal WAN EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO WAN m. S 2 m. S 1 Edge n m. S 1 m. S 2 m. S 1 Edge Platform Network (LAN/ WAN) 3

Geo-Distributed Computing - few use cases 5 G Use case Enterprise (Retail) use case Universal CPE Use case • • • Large Number of sites Computing (Apps across sites) – MEC Multiple tenant applications along with operator CNFs. Workload types - VMs, VNFs, CNAs and Functions (Faa. S) Note: K 8 s is becoming choice of workload orchestrator in each cluster Multi Edge/Cloud computing scale is similar (or even higher) to Hyper-scalers’ scale Now Telcos, MSPs and Enterprises need @scale Orchestration and Automation solutions EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO

Edge Computing – Similar to Cloud-computing, but with some special needs Uniform Developer Experience across Clouds and Edges Resource Constraints (Power, Cost, Space) Edge requires high security assurance (No physical security in far edges) Ease-of-Use (@Scale requirements are higher than the clouds) 5 G Breakout EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Support for all kinds of workloads ( VM, Containers and Serverless functions) Easy migration of workloads among Edges and Clouds Multi Cloud Uniform Networking (Overlay) Converged Edge supporting IT, OT applications & Network functions Optimized infrastructure software Accelerator usage (Hence, awareness without losing platform independence property Platform attestation Confidentiality Multi-tenancy isolation Infrastructure Orchestration (K 8 s Cluster Life cycle management) Multi Cluster Distributed Application Orchestration & LCM 5 G UPF RAN Acceleration 5

E 2 E Edge Stack CI/CD workflows Infrastructure Orchestrator Distributed Edge Orchestrator (Multi Cluster Orchestrator) Multi Cloud Networking Hubs Distributed Applications & Network functions Edge Extensions (Open. NESS) Edge Extensions K 8 s Linux VMs K 8 s Linux Private Cloud K 8 s cluster EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Public Cloud K 8 s Cluster Edge K 8 s Cluster Possibly in hundreds/thousands

Service Orchestrator – Big Picture One Click deployment of complex applications & network services across multiple K 8 s clusters Service Orchestrator Comprehensive Status monitoring of deployed complex applications One Service Orchestrator for both CNFs and Applications Self Service Portal for multiple tenants Comprehensive Analytics platform for Day 2 operations Single pane of glass for Day 0/Day 1/Day 2 configuration of CNFs and Apps. EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO

Needs/Requirements – Preparation Registration of Clusters Service Orchestrator Cluster Provisioning Cluster labels (Example: Cell tower Edge, CO Edge etc. . ) Needed for identifying multiple clusters Cluster specific configuration (Few: ISTIO CA provisioning; Virtual/Provider network preparation) Tenant Provisioning Logical Cluster Provisioning Tenant registration Ability to use tenant specific OAUTH 2 servers for authenticating tenant admins Logical cluster Tenant level isolation via RBAC rules Logical cluster Logical Cluster provisioning across multiple selected clusters Logical Cluster user and permission provisioning EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO

Needs/Requirements – Application deployment design App Onboarding (Complex Apps & Network Services) Service Orchestrator Cluster Provisioning Tenant Provisioning Logical Cluster Provisioning App/Service onboarding Define placement policies Define Placement constraints Multiple deployment profiles to ensure same APP can be instantiated multiple times Placement policies to replicate and distribute workloads across clusters Placement constraints : Affinity and Anti-Affinity; Platform capabilities; Latency; Cost Tenant level isolation via RBAC rules Logical Cluster provisioning across multiple selected clusters Logical cluster EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Logical Cluster user and permission provisioning

Requirements – Workload Customization & Connectivity management No changes to helm charts/K 8 s description of applications Service Orchestrator Cluster Provisioning Tenant Provisioning Logical Cluster Provisioning App/Service onboarding Define placement policies Define Placement constraints Workload Customization Automation of SM; SD-EWAN; firewall/NAT; SLB; DNS Each deployment may have its own customization - Connectivity intent provisioning Enabling inter-micro service communication within or across clusters - Enabling communication to external entitles - With/Without Mutual TLS - Multi Cluster DNS management Dynamic provisioning with LCM of Applications Logical cluster EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Extensible framework to add new capability controllers

Requirements – Operations Continuous monitoring of complex Application (Across clusters, apps and micro-services) Service Orchestrator Cluster Provisioning Tenant Provisioning Logical Cluster Provisioning App/Service onboarding Define placement policies Define Placement constraints Workload Customization Automation of SM; SD-EWAN; firewall/NAT; SLB; DNS Continuous App monitoring Analytics & Closed loop automation Comprehensive report on the application status Day 2 Config Controllers Analytics framework - Metric collection across clusters & apps - Long term central store (Time Series) - Training framework - Closed loop policy management Logical cluster EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Day 2 Configuration of apps/network-functions that are already deployed. Various types of configurations (CR based, RESTful based or Netconf/yang based)

EMCO – Edge Multi Cluster Orchestrator EMCO is an implementation of Service Orchestrator Addressing majority of requirements; Extensible architecture allows new automation requirements EMCO CLI/GUI Cluster Registration Controller Distributed Application scheduler Hardware Platform Aware Controller Distributed Cloud Manager Traffic Connectivity Controller Day 2 Cfg Generic Resource Synchronizer & Status Monitoring Platforms Enterpris e Edges Edge Clouds Network Edges Telco CO Edges Pub/Pvt Clouds • Cluster Registration Controller registers clusters by cluster owners • Distributed Application Scheduler provides simplified, and extensible placement; tenant mgmt; LCM implementation • Hardware Platform Aware Controller enables scheduling with auto-discovery of platform features/ capabilities • Distributed Cloud Manager presents a single logical cloud from multiple edges • Traffic Connectivity controller auto-configure service mesh (ISTIO) and security policy (NAT, firewall), DNS and SLB entities of edges. • Day 2 generic configuration configures Day 2 configuration of any app/network function via templates & configs. • Resource Synchronizer & Monitoring synchronizes resources across multiple edge/cloud platforms and then monitors the status of deployed resources Open source EMCO can be downloaded from here EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO 12

EMCO Integrations Part of these commercial solutions Part of AMCOP solution from Aarna: https: //www. aarnanetworks. com/amcop Two SIs are using this in their solutions Parts of it are in one 5 G deployments Few Blueprints in LFE/Akraino use EMCO for Multi Cluster Orchestration ONAP uses EMCO to Onboard/design network services and deploy on K 8 s clusters ONAP Slicing Orchestrator uses EMCO for Day 2 LCM EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO

EMCO Roadmap 1) Scheduling with awareness of platform-capabilities 2) Service Mesh Orchestration across Kubernetes clusters 3) Upgrades/Updates 4) Generic K 8 s resource customization Centralized deployment of applications & NFs into multiple Kubernetes clusters 1)One-click deployment of “Distributed Applications & Network Functions” across multiple Kubernetes clusters 2) Metrics collection and centralized store 1) Security Orchestration 2) SD-EWAN Orchestration 3) AI Analytics & Closed loop 4) Resiliency with Uber Cadence 5) Day 2 configuration 6) Analytics & Closed loop EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO 14

Thank You!! EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO 15

EMCO Vision Be a comprehensive geo-distributed Cloud native application orchestrator Be a Multi-Party and Multi-Cloud Orchestrator Be an orchestrator for Network services and Enterprise applications Be an orchestrator for convergence of Network services and Enterprise applications Be an orchestrator for Distributed Clouds with Edge-computing EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO 16

EMCO Market Interest Market u. CPE 2. 0 & 3. 0 Use case • LCM of SD-WAN, Security CNFs with u. CPE for Edgecomputing Interest Tier 1 MSP 5 G • LCM of Disaggregated RAN (DU, CU-CP, CU-UP) • LCM of 5 GC with d. UPF • LCM of Private-5 G • Enterprise Edges & LCM of Enterprise applications • Robotic Arm • LCM of backend of applications on PC/Laptop Few Telco interest CNF Vendors interest SIs interest Enterprise Industry 4. 0 Cloud Native PC EDGE MULTI-CLUSTER ORCHESTRATOR | EMCO Good interest 17