Service Fabrics Enable Network Service Virtualization

#SDN #Devops #Cloud You gots to have some management ....

Nick Lippis, who writes the eponymously named Lippis Report, had a fascinating report on the differences between enterprise and service provider environments with respect to network <something> virtualization.

He observes, through a survey of the ONUG (Open Networking User Group) membership, that what the enterprise needs is Network Service Virtualization (NSV), which he and ONUG define as the virtualization of "enterprise appliances, such as firewalls, load balancers, application accelerators, application delivery controllers, Intrusion Protection Systems, WAN optimizers, call managers, etc., instantiated for each application." (Lippis Report 217: It’s Network Service Virtualization in the Enterprise rather than Network Function Virtualization )

This model, which results in a per-application service instance (e.g. "if there are 10 applications that require network services, then each application will be configured with its own instantiation of that service. That is, 10 applications, then 10 NSV firewalls."), is markedly different from the service provider model, network function virtualization (NFV), which instantiates instances based on service needs. For example, a video optimization service may be offered to either application providers or subscribers. Its provisioning and scalability model is based on demand for that service and is not provisioned per-application being optimized.   

According to Lippis, "NSV hopes to present significant capex and opex relief from hardware appliances, as well as an efficient way of applying network services to applications without chaining or tagging packets and rapid automated, on-demand application deployment."

This is a goal that's shared by many emerging technologies like SDN and cloud computing.

Service Fabrics Enable NSV

A service fabric enables network service virtualization by decoupling the service platform from the underlying hardware. This is accomplished via virtualization techniques and, when associated with COTS hardware generally means an industry recognized hypervisor. The use of virtualization implies rapid provisioning and the ability to leverage commoditized hardware. This process can be automated and integrated with other systems, enabling capabilities such as auto-scaling and high availability.

A service fabric goes beyond the cost and time savings realized by service virtualization by embracing programmability at the service instantiation layers. The potential risk of a NSV approach without an accompanying automation and orchestration strategy is service sprawl and its negation of cost and time savings by the need to individually configure by hand each service provisioned. If an application is going to have 10 times the services, it's going to have 10 times the configuration needed. The advantage service providers have is that its services are not driven by the application, but rather by the service definition. In the enterprise, a per-application service implies per-service configuration, tailored to the application's unique needs and requirements.

That could complexify the environment. That's not desirable when one of the goals is to reduce costs, because complexity has the opposite effect.

 

 

The inclusion of flexible, programmatic control and application service planes can mitigate this potential pitfall. A unified framework with appropriate templates, scripting and APIs along with a strong portal-based management system can result in continued operational savings through automation and orchestration. Such a framework also lays the foundation for IT as a Service (or Software Defined Data Center (SDDC) if you prefer), with the goal of enabling application and business stakeholders the ability to self-service applications through simple, automated provisioning and management portals. 

Network Service Virtualization would certainly be a step in the right direction for organizations seeking goals of either SDDC or cloud computing. The abstraction of the resource layer from the service platform and, subsequently, the automation and orchestration of services can create a rapid, cost-effective yet application-driven service infrastructure, ensuring that every application can take advantage of performance, security and reliability services that live at layers 4-7.