Further NFV can also offer support to SDN by providing the essential infrastructure upon which the SDN software can be implemented. Additionally, NFV works in close alignment with SDN objectives to use commodity servers and switches
Key Challenges Affecting SDN - NFV Adoption
Before outlining the Validation Approach, let’s take a look at the major challenges that a robust SDN - NFV blueprint must overcome:
Openness and Standardization
Most traditional network equipment vendors’ approach to Open SDN does not reflect ideas like open code, architecture and adherence to standards like ONF, ETSI, etc.
Migration to generic COTS-based hardware involves the minimizing performance degradation on latency, throughput and processing overheads.
Co-Existence and Compatibility
For existing IT and SDN / NFV executions to converge, the SDN and NFV implementation must co-exist with the operator’s legacy equipments and be compatible with their OSS, BSS, Element and Network Management Systems.
Portability and Interoperability
Interoperability and portability between multiple vendor solutions and standards is crucial for a SDN/NFV orchestration. It establishes the ability of NFV to load and execute virtual appliances in different but standardized data center environments, provided by different vendors for different operations.
Absence of standard tools to measure SDN/NFV functionality and performance on reliability, stability and portability poses a challenge to the adoption of SDN and NFV.
As the industry moves towards cloud deployment, the availability of skilled man power with the ability to design, build, support, validate and manage a new software-centric carrier network over a long time is a major challenge.
The shift from a hardware-focused implementation to a hardware and software based cloud-centric approach poses a major challenge for SDN and NFV validation. SDN and NFV mostly relies on COTS hardware platforms and this implies a fundamental revision of traditional testing methodologies.
Further, there are several new validation challenges which arise due to a plethora of factors - Multi-vendor software, protocols and API inter-operating in COTS environment.
While engineering the validation solutions for SDN and NFV, a number of qualitative and quantitative criteria and dependencies need to be evaluated:
- What kind of server resources - memory & CPU - must be allocated for each function? » Should resources be optimized for specific functions?
- Can the platform ensure composite compatibility and successful co-existence between the SDN-enabled infrastructure and existing legacy networks?
- Where will the NFV be deployed? Will NFV location affect service delivery due to network delays?
- How can SDN and NFV be rendered scalable? Is automation required? And how would this be integrated with the Cloud?
Given the above environment, it is essential to introduce a robust and reliable validation approach for the smooth migration of network functions and programmable and scalable network to the cloud environment.
Integrated Approach to SDN - NFV Validation
As we move proven network functions such as routing, policy, firewall, DPI and many other from physical hardware-based appliances to running on unproven virtualized server platforms, it necessarily calls for the complete re-validation of previously deployed network functions along with Network Function Virtualization Infrastructure (NFVI) and network orchestration components.
The network orchestration may primarily consist of SDN controllers which deliver a complete view of the network and has the flexibility to provision, manage and de-provision network elements according to NFV orchestrator’s instructions.
Typically SDN controllers are connected to cloud orchestration frameworks via appropriate North Bound APIs (e.g. Neutron Module in OpenStack); hence an integrated validation approach is necessary to ensure end-to-end network services. Figure 1.2 depicts the integrated SDN & NFV framework.