|Topic:||Network Functions Virtualisation – Going Small has Big Upside|
|Title:||CTO & SVP, R&D|
Prayson Pate, Chief Technology Officer and Senior Vice President, R&D.
Prayson co-founded Overture and brings more than 28 years of experience developing software and hardware for networking products to the Company.
He began his career in 1983 at FiberLAN, moved to Bell Northern Research (Nortel) and joined Larscom in 1992. He has extensive hardware/software design and management experience, and served as Director of Engineering at Larscom before co-founding Overture. Prayson is active in standards bodies such as the MEF and IETF, and he was chosen to be the co-editor of the pseudowire emulation edge-to-edge (PWE3) requirements and architecture documents (RFCs 3916 and 3985).
Prayson holds a BS in electrical engineering and computer science from Duke University and an MS in electrical and computer engineering from North Carolina State University. He holds nine patents.
Moving to virtualised appliances presents a big opportunity to today’s service providers. They see that replacing physical appliances with software running on servers opens the door to lower capital expenses (capex) and operational expenses (opex) by changing how services are delivered. One opportunity that is not widely discussed is how this lower cost point opens new markets.
Moving to virtualised appliances presents a big opportunity to today’s service providers. They see that replacing physical appliances with software running on servers opens the door to lower capital expenses (capex) and operational expenses (opex) by changing how services are delivered. One opportunity that is not widely discussed is how this lower cost point opens new markets. In particular, there is a huge market for advanced managed services in the small and medium business (SMB) sector that can’t be effectively addressed today. Network Functions Virtualisation (NFV) may be a way to open the door to those small opportunities that make up a big market. See Figure 1.
Figure 1. Network Functions Virtualisation (NFV) may be a way to open the door to those small opportunities that make up a big market.
Service providers have embraced the notion of NFV because of the promise of replacing expensive and closed appliances with open software running on standard servers. By opening the appliances and separating the hardware and software, service providers plan to achieve a number of benefits:
• Lower hardware costs by replacing dedicated appliances with lower cost servers that can be shared.
• Lower operational costs due to not needing to deploy and maintain appliances.
• Support for on-demand, pay-as-you go deployment models.
• Enable innovation by making it easier to develop network functions.
Most discussions on the value of NFV focus on replacing the appliances for today’s services. This is a valuable goal, especially because it is widely believed to be achievable and provides a very attractive return on investment. However, there is a whole new world of opportunities that are opened up by lowering the cost points, accelerating the introduction of new services and enabling innovative mash-ups of today’s disparate service offerings. We’ll take a look at how this could play out, but first let’s see why getting small is an interesting market direction.
ZDnet reported that Comcast CFO Michael Angelakis indicated Comcast has a US$2B business serving companies with 20 or fewer employees and, for companies with 250 to 500 employees, he estimates the market at US$10B or more. As with all market estimates, the accuracy of this estimate can be debated. Even so, it is clear that the opportunity for providing managed services to SMBs is huge.
This SMB market is no secret and it is currently being addressed by service providers. However, the penetration is limited because capex and opex costs are driving prices higher than what the typical business in this sector can pay. What’s driving this cost? In a word: appliances.
Current service models often require a new appliance for each incremental service. This approach adds capex for the purchase of the appliance as well as opex for its installation and maintenance. In addition, the time to service is slowed by the need to deploy an appliance.
Figure 2 shows an example deployment, where an Ethernet Access Device is co-located with multiple appliances.
While the deployment shown looks expensive, the actual situation may be worse. That’s because services are often added incrementally, with a new appliance being delivered with each new service – which includes a delivery by lorry. In short, the value of the service upsell is impacted by the incremental cost of its delivery. The key to lowering costs and broadening the market for a service is to eliminate appliances and associated deployments of lorries.
Lower barriers to deployment
So we get rid of all but the most essential appliances. What does that buy us? As it turns out, a lot. Service providers are now in the position of being able to offer an advanced service at the same capex as that of a basic service, because they only have to deploy a low-cost Ethernet access device at the customer site. The big change is that there is now an opportunity for upsell. Figure 3 shows how physical appliances at the customer premise are replaced by virtual network functions at the central office or point of presence. This virtualisation of network functions opens up all kinds of opportunities for the service provider.
The most obvious of these opportunities is the upsell. An on-net customer with a simple service can now be offered an advanced managed service with almost instantaneous service activation. This short time-to-service is enabled by the fact that the service is delivered by a virtual network function that is spun up on demand, instead of via a physical appliance delivered by lorry.
This benefit is compounded with each additional service that is ordered, giving service providers a strong incentive to get the customer to take as many services as possible. Fortunately, this new virtual model provides a dandy new tool for driving customer demand: Try before you buy.
Because the incremental cost of an additional service is fairly low, the cost of providing a service at no charge is very manageable. Service providers can now use the techniques of discounting and free trial periods to drive demand, all without breaking the bank in terms of corresponding capex and opex.
One of the big benefits of cloud-based virtual network functions is the principal of pay as you go. There is no need for a long-term commitment or initial outlay of cash. The customer can sign up for a service with little or no commitment, and likewise the service provider will incur license fees only for the period of actual usage. This mode of operation facilitates time-limited services, including:
• Services associated with temporary event-based increases in bandwidth such as the World Cup.
• Enhanced services needed for time-limited activities such as VPN and escrow for merger and acquisition activities.
• Seasonal services to support holiday shopping or travel.
Figure 4: Pay as you Grow Services
Each of these situations may be limited in time, but they represent recurring incremental revenue opportunities for service providers.
But, where do the virtual network functions reside?
The answer is, ‘it depends.’ Figure 3 shows an application where the virtual network functions are placed in a central office. This works well for many services such as basic firewall and VPN. However, server resources are more limited and expensive in the central office (CO) than they are in a metro datacenter. So why not host all of the virtual functions there? Because some services have dependencies on latency and throughput that may not be well-suited to the centralised hosting of a CO.
In addition, hosting virtual network functions in the CO may not work in some specific applications. Consider the case of a customer who is very concerned about security. Their threat may include the usual hackers coming in from the internet, but also intrepid data thieves who ‘sniff’ the access link with sophisticated electronic equipment. For these customers, security functions must reside on their site. Likewise, a virtual network function using deep packet inspection for traffic prioritisation on a low speed link must implement its functionality at the customer site. By the time the data gets to the CO, some data has already been discarded in a priority-insensitive manner.
The common theme here is that different services require different implementations. The beauty of virtual network functions is that they allow a single implementation of a network function such as security or routing to be deployed in multiple locations to optimally meet customer requirements.
We have talked a lot about how service providers can grow their revenue by virtualising today’s services thus increasing the addressable market. A complementary approach is to combine various virtual network functions with other cloud-based services to create valuable new offerings for their customers. This combination of currently disparate pieces is facilitated by the use of cloud principles such as open APIs, modern programming tools, and rapid development and deployment of new services. What form might these new services take? Here are some possibilities:
• Network-as-a-service delivered to the end user. By using open APIs, an end user would be able to dynamically configure their network to support their business-critical applications. While small businesses might not utilize this capability directly, it opens the door to third party suppliers to create dynamic new network-based services personalized for the small business.
• Virtual office functions. Since we have virtualized the network functions, why not use them to support the movement to telecommuting and telepresence? Doing so would enable small businesses to lower travel costs and offer valuable benefits to their employees.
• Dynamic private cloud. Today’s end users can get dynamic cloud services over the internet, and they can also get cloud services delivered over static private network connections. How can they get cloud services that are both private and dynamic? One way is to couple dynamic private networking with dynamic cloud services. Such an approach would greatly benefit businesses that are too small to build their own private cloud services.
Strength in numbers
While each small business represents a tiny amount of incremental revenue, the huge number of these businesses means a great opportunity for service providers. However, this opportunity can only be realized by changing how services are delivered to take advantage of benefits of virtual network functions.