What Is Automated Service Deployment?

Automated service deployment refers to the process of automatically deploying services in a centralized manner across the IP and optical layers by coordinating the IP and optical resources. This implementation avoids coordination between different departments and shortens the service deployment time from several months or weeks to less than 1 minute.

In comparison, non-SDN-based networks are often confronted with problems such as long service deployment time and high OPEX.

By centrally managing multi-layer resources and implementing multi-layer path calculation and establishment based on service requirements, the SDN-based IP+optical solution enables automated service deployment, shortening service provisioning time and reducing OPEX.

Prerequisite for Automated Service Deployment — Multi-layer Path Calculation

Multi-layer path calculation is essential to automated service deployment on an IP+optical network. This means that the IP and optical layers must share resource information.

The IP PCE and optical PCE can collect resource information in their respective domains. In addition, the two PCEs can work together to calculate optimal end-to-end service paths. The IP PCE uses PCEP to exchange information with routers and the optical PCE. The optical PCE uses PCEP to exchange information with transport devices and the IP PCE.

As shown in the preceding figure, R1 requests to establish a service path to R4. This service path has the following constraints:

• Bandwidth
• Delay
• SRLG

NOTE:

At the IP layer, R2 and R3 are unreachable to each other. At the optical layer, however, R2 can communicate with R3 over VNT link R2 -> N1 –> N2 ->R3. The VNT links are planned in advance and imported to the IP PCE before path calculation.

After the IP PCE receives the request of establishing a TE link from R1 to R4, the IP PCE, which cannot find an available IP link between R1 and R4, turns to its VNTM module for help. The VNTM module then provides an available optical path R2 -> N1 -> N2 -> R3. The optical PCE calculates an explicit path from N1 to N2 based on delay and SRLG constraints and returns the result to the IP PCE after verifying this path. The IP PCE then requests R2 to establish a GMPLS UNI tunnel R2 -> N1 -> N2 -> R3. After the GMPLS UNI tunnel is established, R1 finds a path from R2 to R4 at the IP layer.

Multi-layer path calculation offers the following benefits to carriers:

• End-to-end path calculation: Unlike single-layer path calculation, multi-layer path calculation is centrally performed by the IP and optical PCEs in collaboration. With the help of the NetMatrix, the SDN controller can provide automatically calculated end-to-end multi-layer paths with few manual interventions.
• SLA assurance: The optical PCE maintains a global view of the optical network, whereas the IP PCE maintains a global view of the IP network. With the help of the VNTM module, the IP PCE can gain a deep understanding of the underlying network, making it possible to deploy services according to SLA requirements. Unlike partial or segmented path calculation, the IP and optical PCEs can take into account all path calculation constraints during path calculation.
• Improved work efficiency: The IP and optical PCEs automatically collect network-wide resource information through PCEP and GMPLS-UNI (RSVP), freeing network operators from heavy workload.

Automated Service Deployment Process

Traditionally, a carrier needs to coordinate multiple departments to deploy a new service, which is a time-consuming. In the SDN-based IP+optical solution, services can be automatically deployed.

After network hardware resources are ready, you can configure service requirements using either the NetMatrix or multi-layer planning tool. The NetMatrix allows you to configure network services and maintain network operating status in real time. With automatic service deployment and path calculation functions, services can be automatically and conveniently deployed end to end.

The automated service deployment process is as follows:

1.         1.The NetMatrix obtains network topology information and basic NE configurations from the U2000.

2.         2.The NetMatrix delivers planned VNT data to the IP PCE.

3.         3.A user configures an L3VPN service through the NetMatrix GUI. 

4.         4.The NetMatrix decomposes the L3VPN service into VPN instance and LSP requirements.

5.         5.The NetMatrix delivers L3VPN configurations to the IP layer and an LSP establishment request to the LSP ingress.

6.         6.The LSP ingress sends an LSP path calculation request to the IP PCE. After path calculation, if the IP PCE finds that the corresponding IP link already exists, the IP PCE returns the path calculation result to the ingress. The ingress then automatically establishes the LSP and deploys the L3VPN service.

7.         7.If the required IP link does not exist, the IP PCE sends a query message to the VNTM to obtain the required VNT link information. 

8.         8.The IP PCE sends a VNT link path calculation request to the optical PCE. After calculation, the optical PCE returns the calculation result, including the PathKey information, to the IP PCE. 

9.         9.The IP PCE triggers the source C node of the GMPLS UNI tunnel corresponding to the VNT link to establish the GMPLS UNI tunnel.

10.     10.The source C node interacts with the source N node using GMPLS UNI to establish the GMPLS UNI tunnel.

11.     11.After the optical path is established, the source C node uses an IGP to instruct the IP PCE to refresh the IP topology. 

12.     12.After the IP PCE sends the LSP path calculation result to the LSP ingress, the ingress automatically establishes the LSP and deploys the L3VPN service.

The SDN-based IP+optical solution provides simple and user-friendly GUIs for automated service deployment. Carriers do not need to care about the network structure or status during new service deployment. The SDN architecture automatically deploys services based on specified requirements, policies, and constraints.