Posted: December 13th, 2016 | Author: Special Contributor | Filed under: Industry Insights | Tags: NFV, TOSCA, virtualization | Comments Off on Putting TOSCA to Work in Real-Case NFV
Comptel Offer Example Templates to the Open Marketplace
There remains no doubt that virtualization is shaking the foundations of the telecommunications industry and is here to stay. As the technology continues to mature and evolve, use cases become more realistic and so do the requirements to represent the service specifications that allow their programmatic utilization and consumption.
Several standards organizations have laid proposals to this purpose and there appears to be consensus that, in the Network Functions Virtualisation (NFV) case at least, TOSCA NFV appears to be positioning itself as the preferred option for operators and vendors alike.
The Topology and Orchestration Specification for Cloud Applications (TOSCA) is a data model standard managed by industry group OASIS that can be used to orchestrate Network Functions Virtualization (NFV) services and applications. TOSCA does this through a collection of information models and templates to orchestrate applications seamlessly across multiple cloud domains which is ideal for network functions that are virtualized and deployed in datacentres.
At the time of this writing, the latest release of the TOSCA NFV Simple Profile dates from mid-March 2016. The document provides a good insight but it lacks practical, consistent examples (e.g. there are a few errors) and examples from additional sources are hard to come by. This applies to the Cloud Service Archive or CSAR (pronounced Cesar) packaging mechanism, in which the specification is conveniently wrapped with all the necessary components.
[Comptel] has taken the information that’s publicly available and created some examples that we would like to share with the broader community.
Use Case: vEPC Core Network CSAR
A basic representation of the Evolved Packet Core (EPC) Network Service Descriptor (NSD) composed of four Virtual Network Functions (VNFs) as shown in Figure 1 below. In addition, every node has a connection to a common management network.
Figure 1: vEPC Architecture & Interfaces
The CSAR file contains metadata, the service templates or specifications, images and the corresponding scripts for the VNFs themselves. Figure 2 shows the structure of the file.
Figure 2: CSAR file structure
As displayed in Figure 3, the metadata file contains in line 4, a pointer to the main driving template, in this case the overall vEPC NSD which will link to the individual nodes (e.g. VNFs) and their relationship and corresponding connectivity.
Figure 3: Contents of metadata file (TOSCA.meta)
The Network Service Design
The NSD provides the global overview (refer to snapshot below) on how the different components (e.g. VNFs, VLs, FGs, etc.) come together. Lines 9 thru 13 point to the VNF templates, in this case for every VNF.
The individual VNFs are described in lines 26, 37, 49 and 58 respectively. They contain a reference to the type, the list of (virtual) networks they are connected to and in those cases where applicable, a declaration of the forwarding graph capabilities (e.g. lines 45, 46 and 47). Additional details on the VNF themselves are contained on their own descriptors (VNFDs) which are shown later.
Next are the details of the external connection points (CPs). These are demarcation points for the NSD as depicted in Figure 4 and they are described in lines 65, 73, 81 and 89.
Figure 4: External Connection Points
Finally, the networks interconnecting the VNFs themselves. In this case, all networks are point-to-point connections (e.g. ELINE) except for the management one, which is shared across all VNFs (ELAN). Every declaration, as seen in lines 97, 108, 113, 119, 125, 131 and 137 indicates the number of network entities attached to them.
The Virtualised Network Function Descriptor
The VNFDs provide details on the specifications of the individual nodes. The vPDN GW descriptor is shown below as a reference. Starting on line 42 the connectivity is described. This VNF requires two computational resources as expressed on line 48 (VDU1 & VDU2). Two of its interfaces (CP21 & CP22) are enabled to support Forwarding Graphs (line 51). In this specific case, four standard transactions types are supported through self-contained scripts: create, configure, stop and delete (line 54). The interfaces and their respective networks can be appreciated in general the topology depicted in Figure 5.
Figure 5: PDNGW_VNF Topology
At the end of the VNFD template are two Forwarding Paths (Line 144 and 151). They represent the incoming and outgoing traffic for the PDN gateway. Figure 6 and Figure 7 provide a visual perspective of the traffic flows they control.
Figure 6: Forwarding Path1 on VNFD
Figure 7: Forwarding Path2 on VNFD
NSD_vEPC.yaml – File contents:
vPDNGW_VNF.yaml – File contents:
The standards provide enough tools to cover the most general of use cases, but we expect to see future updates that can target elements of the service description that represent more complex and realistic scenarios, for instance:
- Quality of Experience (QoE) or in general Quality of Service (QoS) features. There are some brief references in the existing standards but this area requires further development.
- The transactions/interfaces need to support more complex features that can allow them to be referenced and consumed more easily by higher order service orchestration processes.
- Forwarding Graphs should include indications of the traffic types.
Although Comptel has worked out these areas for its own products and specifications, the real value materialises when these specifications become open and seamlessly interchangeable by the different components in the architecture.
Posted: January 27th, 2012 | Author: Leila Heijola | Filed under: Events | Tags: charging, Cisco Live!, cloud, mediation, virtualization | Comments Off on Comptel Prepares for Cisco Live! in London
Next week, the Comptel team is again heading to Cisco Live! in London, and we expect cloud to be a popular point of discussion at the show. Cloud services offer great revenue potential for communications service providers (CSPs), but harnessing that potential requires a comprehensive platform dedicated to a very different kind of business.
Our staff will be on hand to demonstrate how the Comptel Virtualization Charging Solution (VCS) for Cloud is tailored toward this new business model. It re-uses many of Comptel’s existing mediation and charging components that are already deployed with various CSPs worldwide, and enables them to create advanced and flexible charging models in a cloud context (e.g. Software-as-a-Service, Infrastructure-as-a-Service and Storage-as-a-Service).
The VCS acts as the mediation layer between a CSP’s cloud environment and billing system by collecting usage statistics of virtual machines managed via e.g. Vcenter, Xen or Hper-V, and then processing network bandwidth data from e.g. Nexus routers and rating the data according to active subscriptions, and finally delivering rated items for billing-based specified time intervals. We like to call this “concept to cash”.
If you’d like to talk with us about cloud services’ impact on BSS/OSS, how CSPs can best manage their network assets for managing cloud services and our VCS solution, come to the Comptel booth (#E3)—you’ll find us just next to the Cisco Industrial Network Solutions demo area. (Some of the demo areas have less technical names such as Bloodhound—unrelated to K-9s, Bloodhound SSC is the ultimate land speed record car!)
One other fun fact: After Cisco Live! in London is complete, the conference venue, ExCeL London, will play a big role in the 2012 Summer Olympics and Paralympics, serving as the host to seven and six events (or 80 and 74 medals) respectively. As my peer Andrew Gavin wrote during the FIFA World Cup in 2010, the network infrastructure must have greater capacity than those of previous sporting events based on the anticipated increase in traffic demand. This will be key to ensuring a high customer experience for the global audience the Olympics will bring to London. Cisco will be supplying the routing, switching, firewalls and IP telephony to approximately 100 venues across the U.K. to support the summer games, and they will be showcasing that during Cisco Live.
We hope to see you at the event!
Posted: May 12th, 2010 | Author: Bob Machin | Filed under: Behind the Scenes | Tags: BSS, cloud, communications service providers, Comptel, Management World, Nice, OSS, outsourcing, SaaS, virtualization | Comments Off on Q&A: Gareth Senior on Comptel's Cloud Strategy
Cloud services have the potential to create a whole new line of business for communications service providers (CSPs). Technology advances, particularly in virtualization and remote communications, are making these new offerings ever more credible, and customers are increasingly intrigued by the potential not only to save costs but also to reduce risk and increase business flexibility. Meanwhile, businesses and enterprises are looking beyond software as a service (SaaS) and evaluating the benefits of outsourcing substantial parts of their IT environment to the cloud.
But, cloud services are not all bright new dawn and silver lining. Hear what CTO Gareth Senior has to say on the issues these offerings create and how Comptel is applying its OSS solutions to help CSPs cash in on the cloud.
Q: What challenges do CSPs need to overcome, not only to deliver cloud services efficiently—but also to make some profit from them?
A: Moving into cloud services is not just an incremental change to the CSPs’ portfolios; they bring real challenges. Cloud services have much more in common with IT services and outsourcing than traditional communications services, and customers will view their providers rather differently, too. For example, enterprises will need demonstrable reassurance that the CSPs can deliver quality of service, performance, security and regulatory compliance. And of course, since cost reduction is a big driver, they will be extremely price sensitive. So in order to succeed, CSPs will need to deliver services at least as effectively as their customers could—and definitely more cost effectively. A tough challenge!
Q: What impact does Comptel see the sale of cloud services having on OSS/BSS?
A: As the services to be delivered are different than traditional services, it will have an impact on traditional OSS/BSS. Resource management is going to be more IT-orientated than network-focused. Service management will have to take place in a distributed environment, with services being bundled together from different sources and likely to include communications elements and components from third parties. In terms of charging for these services, variants on ‘pay as you go’, combined with ‘pay as you grow’ models, are likely, even in business and enterprise propositions. All of this means that even so-called convergent OSS/BSS could struggle to handle cloud services.
Q: How does the Comptel Dynamic OSS portfolio support CSPs and their end-customers in ‘catching the cloud’?
A: Comptel believes that cloud service providers require a comprehensive ‘concept-to-cash’ and probably even dedicated platform that can handle this very different kind of business. Comptel Dynamic OSS is well suited to this. Our products let CSPs:
- Catalog and manage cloud products and services;
- Manage the resources needed to provide cloud-based services, from servers to communication lines, to applications and more;
- Fulfill customer orders, using automated order management processes;
- Monitor and respond to customer activity using pre-defined policy;
- Collect usage information;
- Rate and charge for usage and events; and
- Provide customer and operational visibility of cloud environments.
Comptel is partaking in a number of cloud activities at Management World 2010 in Nice next week, and invites you to join in the industry-wide discussion on making cloud service possible and profitable. Leave a comment or swing by the company’s tradeshow booth (#21)!