Enterprise vCPE Use Case
Use Case Authors:
China Telecom, China Mobile, Huawei, ZTE, ARM, Accenture
8.28 usecase subcommittee meeting.pptx
E-vcpe presentation material in F2F meeting - 9.25?
Description:
This use case is mainly used for the Enterprise Service Provider (SP) to provide the Enterprise2DC service to the tenants both efficiently and safely by deploying enterprise-based vCPE. Enterprise users could receive high-quality and cheap cloud service based on the operator’s BSS Portal and ONAP. And in here, ONAP is used to instantiate and operate the required vCPE in the access network as a cloud service.
For the application scenarios of E-vCPE use case, it supports site-to-site, site-to-DC and site-to-Internet services to the customers. It not only provides the solution of ThinCPE and vCPE in the edge-DC but also CPE solution as an alternative choice for enterprise customer. The solution of vCPE in the edge-DC is an actual deployment scheme which can provide more value-added service in DC and make it easy for access control.
In this case, ONAP will present its function to instantiate, configure and manage both the overlay connection between site-to-site/DC/Internet and VNFs . There are three alternative ways to set up overlay connection, including VxLAN, IPSec and MPLS. It is a stretch goal to fulfill traffic scheduling in WAN network by means of path calculation and load balancing.
Enterprise vCPE Scenario diagram as shown below:
Enterprise customers would benefit from cloud services and automation either on-premises or in the cloud depending on their branch configurations. This will free the enterprise users from the installation, local configuration and administration of LAN/NAT/WAN/VPN networking.
SPs would benefit from a more flexible platform to provide more kinds of virtual services. To implement the configuration remotely will help SPs reduce complexity of the service deployment and cut down the OPEX. Besides, ONAP can manage a large number of business orders from BSS or user self-service systems, which can improve service efficiency. The order management services include three core functions, such as order resolution and calibration, order execution, and order reporting.
The centralized traffic scheduling and optimization of backbone network is the embodiment of SDN concept in backbone network, it will help operators enhance the management ability of operation and maintenance and improve the efficiency of resource utilization.
Use Case Flows
ThinCPE's request to get its own IP address for its WAN interface
· ThinCPE send DHCP Discovery message to vCPE for WAN IP
· vCPE store the DHCP Discovery request and then send access request to vAAA
· vAAA reply access accept message,vCPE restore the DHCP Discovery request
· ThinCPE send DHCP Discovery message to vCPE for WAN IP
· vCPE send DHCP Discovery message to vDHCP for WAN IP
· vDHCP sends a DHCP Offer with an address its DHCP server assigns to ThinCPE's WAN IP
· ThinCPE sends a DHCP Request with the offered address to vDHCP
· vDHCP sends a DHCP ACK
ThinCPE's request to get its own IP address for its LAN interface inside the VxLAN tunnel
· ThinCPE send DHCP Discovery message to vCPE for LAN IP
· vCPE send DHCP Discovery message to vDHCP for LAN IP
· vDHCP sends a DHCP Offer with an address its DHCP server assigns to ThinCPE's LAN IP
· ThinCPE sends a DHCP Request with the offered address to vDHCP
· vDHCP sends a DHCP ACK
Once ThinCPE has acquired its own IP address for its LAN and WAN interfaces, it could have access to DC.
ONAP collect link utilization and the actual traffic flows in real-time. when the flow reach the WAN side, the ONAP will conduct the optimal path calculation through configuring MPLS TE tunnels, so as to realize the function of traffic scheduling.
Cloud Location | VNFs | Code / Vendor |
Edge | vFW | |
vCPE | ||
vDHCP |
Cloud Location | Devices | Code / Vendor |
Router | ||
Switch | ||
Server | ||
IPRAN |
Description:
PCE (Path Computation and Element) aims to conduct the path calculation and load balancing for the ONAP network. In the current network, there are multiple vendor controllers along with different path calculation method, and through the PCE, ONAP can analyze data collected from PNFs thus acquiring the whole network topology and flow, and then conduct path calculation uniformly for traffic scheduling and load balancing.
Architecture Alignment:
PCE needs two kinds of information for path calculation and load balancing including network topology and network flow information. In the PCE loop, the module first gets the BGP-LS topology information from the Controller and the network utilization along with network flow information from DCEA/Holmes.PCE will conduct path calculation and load balancing by using those two kinds of information.And then the analysis results are sent to event bus for closed loop automation.
Description:
NAI (Network Artificial Intelligence) component provides data analysis including data detection, prescription and prediction for efficient and intellectualized ONAP network management. And NAI aims to provide traffic prescheduling, policy presetting and resources relocation functions based on the historical data collected from DCAE. Through introducing machine/deep learning algorithms and frameworks, NAI fulfills the fast scheduling automatically and with the well trained NAI system, the accuracy and efficiency will significantly increase compared to the current network management.
Architecture Alignment:
NAI consists of a brain system and a Network Telemetry Analytics (NTA) engine. The NTA engine collected historical data from DCAE then send the data to the brain system to train. With the well trained NAI system, based on the real time data collected from DCAE, the NTA engine generates analysis results to the DCAE and event bus regarding real-time scheduling, prescription and prediction. And the analysis results are the root cause to drive the policy run-time for automation operation.
Network Telemetry Analytics (NTA) engine generates analysis results based on rules and the resources relationships from A&AI.
Current Issues:
Users trigger service design:
Report information to DCAE:
NAI Loop:
PCE Loop:
Scaling and Healing:
Functional Platform Requirements | Priority | Basic/stretch goal Default basic goal | Related components |
Connectivity configuration | 1 | SDNC | |
Traffic scheduling | 3 | PCE, NAI, DCAE, SDNC, A&AI | |
Overlay selection | 1 | Portal, SO, SDNC | |
Order management | 2 | Common Service, A&AI, SO | |
VNF onboarding | 1 | ||
Service Design | 1 | ||
Service Composition | 1 | ||
Network Provisioning | 1 | ||
Deployment automation | 1 | ||
Termination automation | 1 | ||
Policy driven/optimal VNF placement | 3 | ||
Performance monitoring and analysis | 2 | ||
Resource dedication | 3 | ||
Control Loops | 2 | ||
Capacity based scaling | 3 | ||
Triggered Healthcheck | 2 | ||
Health monitoring and analysis | 2 | ||
Data collection | 2 | ||
Data Analysis | 2 | ||
Policy based healing | 3 | ||
Policy based scaling | 2 | ||
Configuration audit | 3 | ||
Multi Cloud Support | 2 | ||
Framework for integration with OSS/BSS | 3 | ||
Framework for integration with vendor provided VNFM(if needed) | 1 | ||
Framework for integration with external controller | 1 |
1.portal
Support E-vCPE case lifecycle management actions, monitor service instances status, display FCAPS metrics data in real time, select Overlay connection ways(IpSec/VxLan).
2. SDC
Design vCPE network service;
Design policies (eg. Auto-healing policy, design alarm correlation rules );
Design work flow used by SO/VF-C/SDN-C;
Design E2E service…
3. DCAE
Support statistics collection on the CPE and receipt of events as per the new model in real time ,for example network utilization along with network flow information, jitter etc.
4. Policy
Support to define and execute auto-healing policy rules;
Handling the notifications related to VNF and PNF alarms/events ;
5. A&AI
Inventory model for E-vCPE;
Add support to PNFs
Infrastructure resources ;
Topological information, E2E service lifecycle management, decompose E2E service and talk with VF-C/SDN-C to instantiate services respectively in NFV/SDN domain;
6. SO:
E-vcpe service workflow execution (E2E service lifecycle management, decompose E2E service and talk with VF-C/SDN-C to instantiate services respectively in NFV/SDN domain)
7. SDN-C
Network connection service lifecycle management;
Integrate with 3rd party SDN Controllers to setup MPLS VPN for underlay;
Integrate with 3rd party SDN Controllers to setup overlay connection;
Support current node status information collection and so on.
8. APP-C/VF-C
Network service lifecycle management;
VNFs in E-vCPE usecase lifecycle management;
Integrate with EMS to collect FCAPS metrics data of VNF level and do Application configuration;
Integrate with Multi-VIM to do virtual resource management;
9. PCE
Integrate with controller to get the BGP-LS topology
Integrate with DCEA/Holmes to get network flow information
Conduct path calculation and load balancing
10. NAI
Data detection, prescription and prediction;
Provide traffic pre-scheduling, policy pre-setting and resources
relocation functions;
11. Order Management
Take orders from BSS for management: order storage, decomposition, execution status management, completion report, etc
Work Item | ONAP Member |
Modeling | |
SDC | |
SO | |
SDNC | |
DCAE | |
APPC/VF-C | |
A&AI | |
Policy |