The Linux Foundation Is Changing The Fabric Of Networking (forbes.com) 31
Will Townsend, a senior analyst at Moor Insights & Strategy research firm, writes: As it relates to networking, the Linux Foundation is currently focused on a number of projects that are bringing top networking vendors, operators, service providers, and users together. Among the top initiatives are the Open Network Automation Platform (ONAP) and Data Plane Development Kit (DPDK). In this article, I would like to dive into both of these initiatives and share my perspective on how each is transforming the nature of networking [Editor's note: the website may have auto-playing videos; an alternative link was not available].
It makes sense that ONAP's releases are named after global cities, considering the platform's growing global footprint. ONAP is aimed at bringing real-time automation and orchestration to both physical and virtualized network functions. The first release in the fall of 2017, named Amsterdam, delivered a unified architecture for providing closed-loop networking automation. The underlying framework ensured a level of modularity to facilitate future functionality as well as standards harmonization and critical upstream partner collaboration. Initial use cases centered on Voice Over LTE (VoLTE) services as well as Virtualized Consumer Premise Equipment (vCPE). Both are extremely cost disruptive from a deployment and management perspective and deliver enhanced service provider agility. What I find extremely compelling is that Amsterdam was only an eight-month development cycle from start to release. That's an amazing feat even in the fast-paced technology industry.
[...] DPDK was an effort initially led by Intel at its inception nearly eight years ago, but became a part of the Linux Foundation back in 2017. At a high level, the technology accelerates packet processing workloads running on a variety of CPU architectures. DPDK is aimed at improving overall network performance, delivering enhanced encryption for improved security and optimizing lower latency applications that require lightning-fast response time. The transformative power of 5G networks lies in their potential to deliver low latency for applications such as augmented/virtual reality and self-driving cars -- DPDK will further extend that performance for next-generation wireless wide area networks. I had the opportunity recently to speak to project chair Jim St. Leger after the fifth DPDK release, and I was impressed with the depth and breadth of the open source project. Over 25 companies and 160 technologists are involved in advancing the effort. With the proliferation of data, cord cutting at home, and growing consumption of video over wired and wireless networks, high-quality compression techniques will dramatically improve performance and reliability. DPDK appears to be poised to contribute significantly to that effort.
It makes sense that ONAP's releases are named after global cities, considering the platform's growing global footprint. ONAP is aimed at bringing real-time automation and orchestration to both physical and virtualized network functions. The first release in the fall of 2017, named Amsterdam, delivered a unified architecture for providing closed-loop networking automation. The underlying framework ensured a level of modularity to facilitate future functionality as well as standards harmonization and critical upstream partner collaboration. Initial use cases centered on Voice Over LTE (VoLTE) services as well as Virtualized Consumer Premise Equipment (vCPE). Both are extremely cost disruptive from a deployment and management perspective and deliver enhanced service provider agility. What I find extremely compelling is that Amsterdam was only an eight-month development cycle from start to release. That's an amazing feat even in the fast-paced technology industry.
[...] DPDK was an effort initially led by Intel at its inception nearly eight years ago, but became a part of the Linux Foundation back in 2017. At a high level, the technology accelerates packet processing workloads running on a variety of CPU architectures. DPDK is aimed at improving overall network performance, delivering enhanced encryption for improved security and optimizing lower latency applications that require lightning-fast response time. The transformative power of 5G networks lies in their potential to deliver low latency for applications such as augmented/virtual reality and self-driving cars -- DPDK will further extend that performance for next-generation wireless wide area networks. I had the opportunity recently to speak to project chair Jim St. Leger after the fifth DPDK release, and I was impressed with the depth and breadth of the open source project. Over 25 companies and 160 technologists are involved in advancing the effort. With the proliferation of data, cord cutting at home, and growing consumption of video over wired and wireless networks, high-quality compression techniques will dramatically improve performance and reliability. DPDK appears to be poised to contribute significantly to that effort.
Buzzword bingo! (Score:5, Funny)
The underlying framework ensured a level of modularity to facilitate future functionality as well as standards harmonization and critical upstream partner collaboration.
Somewhere in a dark cubicle a marketing manager just creamed his pants after reading this.
Re:Buzzword bingo! (Score:4, Insightful)
It's so bad that this could be an article of The Onion.
Half a page of 'summary' and I still haven't the foggiest idea what the hell ONAP and DPDK are.
Re:Buzzword bingo! (Score:5, Funny)
Half a page of 'summary' and I still haven't the foggiest idea what the hell ONAP and DPDK are.
They're concordant multi-faceted paradigm changing solutions to ensure accelerating velocity for OEM vendors and end user implementations of future requirement conditions within a broader flowing service market... Duh!
Almost like git (Score:2)
They're concordant multi-faceted paradigm changing solutions to ensure accelerating velocity for OEM vendors and end user implementations of future requirement conditions within a broader flowing service market... Duh!
So, it's a bit like the "Git gets easier once you get the basic idea that..." parody.
Re:Buzzword bingo! (Score:4, Informative)
Re: (Score:2, Insightful)
And as someone who runs an extensive OvS/KVM deployment utilizing DPDK with said intel fiber nics, I think the whole situation is still best summed up with "Meh." It's not really any better, it's just another layer of BS piled up on top.
with Intel meltdown! (Score:2)
with Intel meltdown!
Re: (Score:2)
Or a Weird Al song...
https://www.youtube.com/watch?... [youtube.com]
Re: (Score:1)
LOL ... I came to say the same thing ... The underlying framework ensured a level of modularity to facilitate future functionality as well as standards harmonization and critical upstream partner collaboration ... Both are extremely cost disruptive from a deployment and management perspective and deliver enhanced service provider agility ... and allow us to leverage the synergies and optimise capital returns and maximise shareholder value.
It's fucking buzzwords, all the way down.
Re: (Score:2)
The article summary reads like a Dilbert comic.
Advertisement for nerds (Score:1)
Re:Advertisement for nerds (Score:4, Insightful)
What msmash probably doesn't realize is that just because an article has the Forbes logo on it doesn't mean it was written, commissioned, or edited by Forbes staff. Like an increasing number of publications, Forbes has a contributed content program [poynter.org] where anyone who "meets their qualifications" (heh heh) can sign up to publish articles on the site. Forbes claims to "vet" the content, but I've never seen them object to anything. What Forbes gets out of it is free content. What do they care if it's garbage?
Oh wait ... you were thinking /. editors are supposed to weed out crap like this. You must be new here.
Don't have to change it (Score:2)
Just wash it once in a while.
A whole lot of FUD (Score:2)
Get ready, you're going to see a whole lot more of this crap. Basically they want SDN but on a whole new level. VM's as you know it are no longer about isolation but how much they can squeeze on a box and more importantly, the ability for the host to see into the VM. Especially in DPDK.
To be clear, from a performance standpoint, writting directly into a vms buffers IS "faster", the problem is there's zero security in it. Also doesn't help the network stack sucks out of the box even in Linux.
Hardware