Lean NFV was unveiled publicly
at the
Linux Foundation's Open Networking Summit 2019 in April with Sylvia Ratnasamy
(UC Berkeley and Nefeli Networks), Scott Shenker (UC Berkeley and Nefeli
Networks) and Constantine Polychronopoulous (VMware). We've covered a lot of
news about NFV at VMblog in the last few years, and were interested in learning
more about how this movement aims to cut network clutter, so we appreciate Sylvia taking
the time to give us the inside perspective.
VMblog: To kick things off, perhaps you can explain a little more about Lean NFV?
Sylvia Ratnasamy: Lean NFV is an open
architecture approach that gives VNF vendors, orchestration developers, and
network operators a simpler way to approach and deploy NFV. It is designed to
be minimally invasive and can integrate with existing NFV infrastructure
platforms as well as ongoing orchestration initiatives.
At
the same time, Lean NFV does not dictate a canonical implementation. It
provides an overall framework consisting of a number of components and also
defines the interaction points between these components. At the
component-level, implementers are free to innovate.
VMblog: And why Lean NFV?
Ratnasamy: It has been over six
years since the publication of the foundational NFV white paper. However, the
promise of NFV has still been unfulfilled. Much of this stems from the rigid
and complex framework we put in place early on. Based on our collective experience
over the last six years, we better understand how we can streamline
deployments. It's time to look at alternative approaches and take a simpler
approach to help accelerate NFV adoption.
VMblog: Who invented Lean NFV?
Ratnasamy: Lean NFV wasn't
"invented" but is the result of discussions and collaboration between some
industry veterans in the virtualization and networking space. It comes from
real-world experience in deploying NFV.
VMblog: And who makes up Lean NFV?
Ratnasamy: Lean
NFV is comprised of industry veterans who have been intimately involved with
virtualization efforts in telco and cloud data centers. We are a collection of
community members who have come together to help improve the success of NFV. To
see a list of founding signatories, check out the Lean NFV white paper at LeanNFV.org.
VMblog: For the VMblog audience reading this, how can they get
involved with the Lean NFV community?
Ratnasamy: The best way is to email info@leannfv.org and one of the community
members will reach out.
VMblog: What about getting a copy of Lean NFV?
Ratnasamy: Lean NFV is an open
architecture, not an open-source project. We will be working as a community to
provide example implementations of Lean NFV using open-source components. Until
those implementations become available, there's no Lean NFV distribution that
can be downloaded. In the meantime, we are also working on reference design documents and
plan to publish the first of those in the near future.
VMblog: Does Lean NFV replace ONAP and OSM?
Ratnasamy: No,
Lean NFV is not intended to replace existing orchestration systems. We intend
to have Lean NFV integrate with existing orchestration platforms like ONAP. In
particular, we would anticipate that ONAP will generally be used as a
multi-domain or higher-level orchestrator, and a Lean NFV system could be
orchestrated by ONAP.
VMblog: Does it work with or replace anything within the OpenStack framework?
Ratnasamy: The
Lean NFV framework can operate in environments that have or prefer OpenStack as
a VIM (virtualized infrastructure manager). At the same time, the Lean NFV architecture
does not dictate the use of a VIM. A bare metal platform with a resource
orchestration solution also conforms to the Lean NFV architecture.
VMblog: Finally, what do you predict for
the future of Lean NFV?
Ratnasamy: We're seeing strong traction from members of the community
already and we have more signatories and contributors to the effort, which
we'll be announcing soon. It's clear that there's a latent demand for an
initiative that can help accelerate NFV deployments today. I see Lean NFV
gaining traction with the community coming together to further flesh out the
architecture, and open source development to demonstrate the concepts. I also
expect proof-of-concept deployments with the Lean NFV architecture over the
course of the next 6 months, where we can get additional learning about how
best to evolve and extend the open architecture.
##
About Sylvia Ratnasamy
Sylvia Ratnasamy is co-founder and CTO of Nefeli Networks. She also serves as an associate professor of computer science at the University of California, Berkeley, where her work focuses on the design and implementation of networked systems. She is best known as one of the inventors of the distributed hash table (DHT), a technology that has been key to scaling distributed applications and databases in the cloud.
In recent years, Ratnasamy has focused her research on programmable networks including the RouteBricks software router and pioneering work in Network Functions Virtualization (NFV). Previously, Ratnasamy was a lead researcher at Intel Labs.
She received her PhD in computer science from U.C. Berkeley and a bachelors in computer engineering from the University of Pune, India. She is a recipient of the ACM Grace Murray Hopper award, the ACM SIGCOMM Test-of-Time award, the ACM SIGCOMM Rising Star award and the Alfred P. Sloan research fellowship.