As telecom and enterprise network operators search for alternative solutions to expensive, proprietary, and static traditional network solutions, the rise of Network Functions Virtualization (NFV) has become unstoppable. NFV simplifies the architecture of physical networks by allowing telcos to virtualize major elements of their networks, replacing dedicated network appliances. Making the move to virtualized infrastructures allows operators to increase networking agility and innovation while decreasing costs…a win-win for businesses with limited communications budgets.
For NFV environments to succeed, however, operators require fast-evolving infrastructure resource management–something that the OpenStack platform can provide. A free and open-source platform for cloud computing, OpenStack is playing a crucial role in the network transformation of today’s telecommunication operators and large enterprises. In fact, according to a survey by Heavy Reading, 85.8 percent of telecom respondents consider OpenStack to be essential to their success. This is due to the plethora of benefits it brings, such as cost savings, flexibility, and more rapid virtualization of their data centers.
Sourceforge recently caught up with Boris Renski, the Chief Marketing Officer and Co-Founder of Mirantis, a leading commercial distributor of OpenStack, to help further understand the subject of NFV and the value it offers to the telco industry. Renski also contributed his thoughts on the Mirantis Cloud Platform, the company’s comprehensive private cloud software stack, and how it helps service providers accelerate their NFV roadmap execution.
Defining NFV
Boris Renski, CMO and Co-Founder of Mirantis
With the cloud and cloud-based technologies picking up speed in the communications market, new technologies are emerging that can enable companies to power up their networking infrastructure like never before. With this increase, however, comes an abundance of new terms and acronyms, which can also lead to confusion, especially for non-technical business people. The terms Virtual Network Functions (VNF) and Network Functions Virtualization (NFV), for instance, are sometimes used interchangeably by networking professionals. Renski mentioned that while these terms are related, they do not actually mean the same thing, which can often cause confusion.
“The simplest way to say it is that VNFs are how you implement NFV,” Renksi explained. “‘Network Function’ is a generic term for networking elements that perform a specific function or role. Classic examples of these network functions include NAT, firewalls, DNS, routers, and caches. Historically, these functions were implemented as physical devices running specialized networking software. These physical implementations are called PNFs.”
Most PNFs are provided by vendors. Any time a new function is needed, a truck roll is required to install new hardware. As service providers began exploring ways to reduce costs and accelerate the speed of deployments, they began to use network function virtualization, which decouples network functions from the specialized hardware that used to be required and moves them to virtual servers running on commodity hardware. Any network function that’s running on some form of a virtual server is a VNF, or Virtualized Network Function, as opposed to a PNF, or Physical Network Function.
“So basically, NFV is an approach that moves PNFs to VNFs and lets you move from vendor-specific proprietary hardware to commodity “off-the-shelf” hardware,” stated Renski.
The Benefits of NFV
According to Renski, the main advantages of the NFV approach to telecommunication service providers are reduced total cost of ownership and agility.
The NFV approach reduces a company’s total cost of ownership in several ways. “For example, migration from vendor proprietary physical hardware to commodity ‘off the shelf’ hardware can save you a lot of money to start with, but it also gives you the flexibility to customize your hardware profile for specific needs by adding or removing servers pretty easily,” shared Renski.
“You can also run multiple VNFs on the same physical server,” he explained. “Disaggregating the stack means you can move from vendor-specific vertical solutions to an integration of multiple independent components provided by different vendors, or even better, by open-source communities, so you can select modules that are the most cost and/or function-effective. It’s all about separating physical hardware provisioning cycles from network function provisioning, so you can do better expense planning.
In regards to agility, NFV enables businesses to expand their opportunities because VNFs can easily be provisioned in only a few minutes with just a few clicks, Renski advised. “You can even automate highly complex E2E network services, so you’re reducing the time from idea-to-cash from weeks or months to days or even hours,” he said. “The important thing is that all of what used to be physical wiring is now virtual, besides being able to scale network functions up and down by just provisioning and deprovisioning VMs, you can configure and reconfigure complex setups quickly using SDN You can also meet tighter SLAs, because you can just create a new VNF as a replacement if one fails.”
“So between decreased cost and the minimized cost of failure for new services, you wind up with this ‘risk-friendly’ network that lets you compete with the likes of Google and Facebook around the world,” he added.
Barriers to NFV Adoption
NFV is the goal for many businesses…but like any evolving technology, there remains a few barriers to NFV adoption. Renski identifies the following as some of the biggest hurdles to NFV:
- NFV ecosystem not developing fast enough. Because of this slow development, enterprises often get various proprietary solutions; and oftentimes, none of these fully meet operator requirements. This is a problem because unlike physical legacy networks, NFV requires interoperability between various components, even at the control and user panels; and because a common set of NFV interface specifications and process flow definitions do not exist yet, this severely limits the growth of ecosystem, even though there are efforts to solve this problem with projects such as OPNFV.
- Unrealistic expectations. NFV offers the critical benefits of agility and lower CapEx and OpEx, but this has led to the false notion that an operator can realize and utilize these benefits from day one; when in reality, an initial investment will be much higher than that of a legacy network. Operators need to create more realistic business use cases that highlight the initial investment being offset by revenue acceleration due to agility in the short term and reduction in CapEx/OpEx in the long term.
- Cultural issues. Traditionally, network and IT were separate organizational silos, and the operations/governance of either organization did not adversely impact the other. With NFV, IT and network have essentially become one, resulting in huge challenges to existing organizational structures. Your organization has to make a change and adopt the DevOps approach to network operations if you want to succeed with NFV. Mirantis is seeing this shift, but it’s a slow shift.
Accelerating NFV Deployments with the Mirantis Cloud Platform
The Mirantis Cloud Platform (MCP) is a comprehensive private cloud software stack that includes leading open source software such as OpenStack and Kubernetes. “MCP addresses the NFV Infrastructure (NFVi) layer — except for the hardware, of course — and the Virtual Infrastructure Management (VIM) layer of the ETSI NFV reference architecture, so as an operator, MCP helps you accelerate your NFV roadmap execution,” shared Renski
The MCP covers the essential NFV requirements, including:
- Universal infrastructure: Operators are aiming at creating a universal NFV infrastructure capable of hosting varying types of network functions. MCP provides the tools to fine tune NFV configurations to fit the needs of different types of VNFs from different vendors.
- Programmable Network: MCP is implemented using the ‘infrastructure as code’ paradigm to eliminate time-consuming upgrades to network infrastructure, which can otherwise take months or even years to achieve.
- Infrastructure visibility: MCP provides complete visibility into the NFVI and VIM layers of NFV in a single pane for monitoring, faults, logging etc.
In July, the company announced a series of NFV-focused updates to MCP, which is available for customers to consume via the DriveTrain toolchain. “The feedback from our clients has been spectacular. In fact, Reliance Jio, which has the largest all IP network in the world, just spoke at the OpenStack Summit in Sydney to talk about the thousands of servers that make up their Mirantis OpenStack cloud, and how they think they might even reach one million cores in the next year or so,” said Renski. “That’s some serious scale for an NFV-centered cloud, and we’re very proud to be behind that.”
So, What’s Next in the NFV Space?
Renski offered some insights into the future of NFV. Two of the things Renski believes will come up in the not too distant future in regards to NFV are operator alignment on automation of NFV operations and cloud-native VNFs.
- Operator alignment on automation of NFV operations. “You can already see this in the ONAP and OPNFV open source communities,” said Renski. “They’re both a new kind of open source project, in that what they produce isn’t actually a deployable product, but rather requirements for an alignment on operationalization and automation of NFV in production environments.”
- Cloud native VNFs. Some of the newer VNFs are being developed as actual cloud-native applications, as opposed to traditional applications just moved to the cloud, so they’re bringing true virtualization and cloudification to telco networks,” he revealed. “When operators have cloud-native options for VNFs, you’ll see the change from physical network functions to NFV really start to snowball, until everything is virtual.”
About Mirantis
Mirantis is a Sunnyvale, California-based company dedicated to converting open source innovation into customer value. Mirantis specializes in delivering open cloud infrastructure to top enterprises using OpenStack, Kubernetes and related open source technologies. To date, Mirantis has helped over 200 enterprises build and operate some of the largest open clouds in the world. Its customers include iconic brands such as AT&T, Comcast, Shenzhen Stock Exchange, eBay, Wells Fargo Bank and Volkswagen.