Archive

Category Archives for "Cumulus Networks Blog"

Networking Field Day at Cumulus Networks – Recap

Networking Field Day is not your typical event; it is an interactive forum where vendors present to selected networking industry experts.

Feb 11, 2015 was our first Networking Field Day at Cumulus Networks and as many have commented, it was “very informational” and an “excellent” presentation. So, for those of you that could not make it, here is the recap with the recordings.

Being an atypical event, #NFD9 had to start off differently with…. a grandiose entrance from the delegates in a limo and some first-class espressos powered by the Cumulus Networks team before getting off to an exciting 2hours forum.

Networking Field Day - Cumulus Networks tweets
Tweets from NFD9

Cumulus Networks Overview

Kicking off the session was JR Rivers (@JRCumulus), Co-founder and CEO of Cumulus Networks for a straight conversation and great anecdotes on Cumulus Networks and what makes Cumulus Linux different. “Cumulus Linux is fundamentally & unashamedly Linux!” What sets Cumulus Networks apart from others? The responsibility to take patches upstream, the real partnerships, being an enabler rather than a gatekeeper.

Networking Field Day at Cumulus Networks

Data Center Architectures

What architectures are best suited for today’s date center network? And how does Cumulus Networks make it easier to build networks? Dinesh Dutt (@ddcumulus), Chief Continue reading

Opening the Black Box – Linux Network OS for Bare Metal Switches

Transform the network from gatekeeper to enabler with a Linux Network OS platform for switches that enables automation, feature velocity and innovation on par with servers.

“The only way to deal with an unfree world is to become so absolutely free that your very existence is an act of rebellion.” – Albert Camus

In my role as a Customer Solutions Engineer (affectionately known as CSE) at Cumulus Networks, I am on the frontlines discussing customer requirements, use cases and networking architectures. A frequent question that customers ask me is “what can an open network operating system (OS) do for me?”

Most customers have lived in the world of black boxes where the OS and hardware are vertically integrated and your vendor keeps you in a sandbox that controls what you can and cannot do. In the black box world, if you want a new feature, application or a different operational model, you have to request it from your account team and wait while the vendor decides if your use case is important enough or you are a big enough customer.

The idea of having direct access to the different operational aspects of the OS is a foreign concept Continue reading

Testing Cumulus Linux with Serverspec

As configuration management software has matured, engineers have recognized the importance of testing and validation. A comprehensive set of tools has been been developed around configuration management software such as Chef, Puppet, Ansible and others that allow you to test your configuration management scripts.

One of these tools is Serverspec, which is an RSpec testing framework for checking that servers are configured correctly by testing their actual state.

Serverspec can execute its tests on a remote host (such as a Cumulus Linux switch) via. SSH. The tests express how the system should be configured and Serverspec will test that the current system configuration meets those expectations.

Using Serverspec to validate your switch configuration means that you can make changes to your configuration management scripts and be confident that the changes have been applied correctly and worked as intended.

Getting started

Because Serverspec natively supports Cumulus Linux, all you have to do is install Serverspec and create your tests. If you’ve never used Serverspec before, the serverspec-init command will create an example called sample_spec.rb. Adding your own files is very simple; just make sure the filename ends in _spec.rb and the first line of the file is require 'spec_helper'

Writing tests

Individual Serverspec tests are Continue reading

My First Week at Cumulus Networks

From the first time I spoke with someone at Cumulus Networks, I realized I’d come across something spectacular. My interviews started with a “screening” call from Nat Morris, but it didn’t start with any of those awkward “tell me about yourself” ice-breaker questions. Nat immediately began the conversation with a detailed description of the Cumulus Workbench, his plans for where to take it, and what I thought about them. For someone like me, being able to talk candidly about a concrete project made it easy for me to see myself in the position, and I felt like I was being treated as if I were already a part of the company. This conversation left such a powerful impression that it literally carried me through the entire interview process.

I started on the day before the RDU team moved from the Apex office to the new office in Cary. All of the furniture was gone, everyone was huddled at folding card tables, and I ended up sitting on the floor in another room since we didn’t have enough chairs even after borrowing some from the restaurant next door! The team assured me that it wasn’t a bait-and-switch by showing me Continue reading

Networking Field Day 9: Fresh Perspectives on Open Networking

The community-selected delegates of Networking Field Day 9 #NFD9 are coming to Cumulus Networks on February 11, 12:30pm-2:30pm PST, to learn more about open networking – and it’s going to be live streamed for everyone, so mark your calendar.

We’ve assembled a team of industry experts from Cumulus Networks to discuss with 13 NFD9 delegates data center evolution and what makes us different.

The topics to be presented are:

  • Overview and what makes Cumulus Linux compelling JR Rivers, co-founder and CEO
  • Data center architectures and how we make it easier to build networks Dinesh Dutt, Chief Scientist
  • Data center network operations and how we make it easier to manage networks Matt Peterson, Office of the CTO
  • Real world examples, from a network admin to a cloud admin David Sinn, Customer Solutions Engineer

Networking Field Day is not just a one-way presentation; it brings us together with independent thought leaders to share information and opinions in a presentation and discussion format. If you’re a networking technologist, this is a great opportunity for you to join the live stream and participate in the discussions on Twitter #NFD9 and @CumulusNetworks. If you’ve heard about open Continue reading

Open Networking: From Concept to Reality

When I first heard about Cumulus Networks in August 2013, I thought, “what’s the catch?” Now, after a year of working with companies deploying Cumulus Linux based switches in their production environments, it turns out there is really no catch. Open networking is for real.

The way you buy a server is now the way you can buy a switch

An open ecosystem has supported the server business for many years. One can build servers with components from various suppliers and run their choice of operating system. This same concept for the networking world, now called “open networking” for the disaggregated model of switch hardware and software, has long been on many wish lists.

The good news: this concept is a reality now, thanks to companies like Cumulus Networks whose OS, Cumulus Linux, is a Debian based distribution that is the OS for open networking on bare metal switches.

No License Gotchas

With Cumulus Linux, there are no additional or “enhanced” license fees akin to what traditional vendors have charged for years. The yearly renewal license fees cost the same each year – not a penny more. The yearly or multi-year license (option) can be ported from one switch Continue reading

Open Networking is the New Normal

The data center is in a constant state of transition. What was once home for rows upon rows of propriety and often siloed equipment based on closed-architecture designs, the modern day data center is now filled with white box solutions serving various functions but working in a harmonious or converged manner.

Several key factors are driving the change to white box or open hardware – ROI, flexibility and customizability of design, ease of implementation, and the avoidance of vendor lock-in along with the high price-tag it can bring. The rise of white box hardware started with servers and storage, and now a movement towards the adoption of open networking has gained quite a bit of traction. The

Open Compute Project (OCP) movement is driving creation of bare metal switches, such as Open Switches, that are designed to be open and disaggregated. This white box model for switching enables users to deploy, monitor, and manage networking alongside servers and storage at a much lower price-point than a traditional network switch.

Scaled Networking Simplified

With a white box switch, the OS layer is decoupled from the hardware itself which allows users to independently select the best-of-breed components and networking software stack Continue reading

Open Networking Accelerated with Help from Experts

Any major technology shift in the data center, such as open networking, is about much more than just swapping out today’s hardware and software for the better, faster, more affordable next generation products.

It’s just as much about people and their processes – and the learning curve required for accelerated success as the new technology takes hold.

Open networking is at this stage now. It’s broadly seen as an inevitable tech trend as data center architectures change to a software-driven model to resolve challenges around agility, scalability and cost. The trend is reminiscent of the transformation of the server infrastructure that gave customers choice, better tools, and TCO improvements – and that allowed data center system admins to develop new technical skills along the way.

2015 is the year when open networking will expand from big cloud operators and early adopters to organizations of all sizes. The technology is proven, and now there’s support for layer 2, layer 3 and layer 3 overlay architectures.

So what does that mean to the people who keep all the world’s data centers running? It’s an opportunity to expand their technical horizons (and their contributions to the business) by learning how to leverage open Continue reading

Bare Metal For the Rest of Us

Bare metal switching technology fueled the world’s big cloud data centers, with a simplified supply chain as the economic driver. Now the technology and economics are hitting the mainstream in open networking solutions for the rest of us.

Bare metal servers changed the world of compute. The same thing is happening in networking: bare metal switches are the foundation for an inevitable shift in data center networking. The movement is called “open networking” but at it’s core, it’s really just about great physical networks with the additional benefits of a rich ecosystem, broad support for automation and monitoring tools, and improved economics.

The big players in bare metal switching – including Accton and Quanta – leverage best-in-class components to deliver high performance switches, and they do it fast. Add an OS (more on that later) and you’ve got a disaggregated model that enables networking that fulfills the vision of the software-defined data center. If SDDC doesn’t matter so much to you, then think of this: bare metal may fulfill your vision of a more affordable, manageable network that gives you the time and funding for more projects that really help your business.

(For comparisons on pricing and shipping speed versus Continue reading

Juniper OCX – Welcome to the Revolution

Early December 2014, Juniper announced their OCX products that are focused on open, disaggregated networking systems.  As one of the instigators of the revolution, it will be intriguing to see which side Juniper is really on.

While competing with Juniper will be interesting, we’re happy to see them recognize the customer drive towards Open Networking.  Juniper indicates that they are joining the ranks of start-ups like Cumulus Networks and industry leaders such as Dell in this inevitable industry transition… avoiding the “head in the sand” perspective maintained by some other networking vendors.

There were four main sources of information as part of the announcement.

Initial reading shows us a focus very aligned with Open Networking. They say things like…

Juniper announced the OCX1100 that combines … Junos® operating system with Open Compute Project (OCP) enabled hardware

Let me say that again: Customers will have the ability to remove Junos and deploy another vendor’s operating system

To some not familiar with Juniper, news that we are embracing an open hardware design might sound counterintuitive in that anything “open” is not aligned with our strategy. On the contrary, Juniper has always embraced open architectures and open Continue reading

Open Networking Has Arrived

“My servers run on Linux. My team knows how to manage Linux servers and networks. It just makes sense for my switches to run on Linux too.” 

What most people don’t know is that many high-end network switches already run on Linux.

Switches from Cisco®, Extreme Networks® and Arista® use Linux to run their switch hardware (the operating system is hidden behind abstractions and APIs). As well, most of these share the same switching silicon products from Broadcom® and Intel®.

We are in the midst of a major transformation in networking. Innovation from companies like Cumulus Networks® and Edge-Core® are leading the way, disrupting the way new networks are deployed and old networks are upgraded.

In my role as head of product engineering at Tuangru, almost every small-to-mid size hosting service provider I talk to is considering open networking. Why? Because it just makes sense.

Open network hardware is more affordable and easy to acquire. The Linux software is familiar and, in most cases, admins prefer it over the next CLI and syntax versions available.

The rise of DevOps and cloud technologies like OpenStack are driving higher levels of automation and uniformity. Continue reading

The Proof Is in The Facebook Data Center Pudding

In case you missed it, Wired just exposed the elephant in the room with last week’s article on the next generation Facebook data center.

For years, anyone who’s had to build out or run a network has handed over large sums of money to the networking hardware titans, without the freedom to choose what to run on that hardware. But I’m sure if you’re someone who placed one of those orders, the thought crossed your mind if this was always going to be the norm.

Every time before you clicked or signed on that dotted line, you wondered whether it’s worth buying from the incumbents and playing in their locked-in world. Maybe deep down you had some burning desire to break away, but were afraid to stray from the blue chip way of life.

I feel your pain and it’s okay because we all want to maximize the value of our dollar. That’s why we all shop for the best choice and at the best price point; otherwise, we will just wait and buy another day.

I mean, you have the freedom to buy the servers you want, so why not have the freedom to buy the network gear that Continue reading

Using New Relic Server Monitoring With Cumulus Linux

One of the most visible trends on the Web today is the “SaaS-ification” of the enterprise. Major productivity functions like email and calendaring, customer relationship management (CRM), and IT systems management are gaining greater value by being deployed as cloud-based services. IT and systems monitoring companies like New Relic are thriving in the cloud as well.

Another major trend, one that Cumulus Networks is at the forefront of, is the transformation of the “switch as a server.” If you aren’t familiar, check out Cumulus Networks engineer Leslie Carr’s excellent PuppetConf 2014 presentation. Since Cumulus Linux supports Debian-based packages out of the box, we decided to take New Relic’s Server Monitoring product for a spin. We wanted to see how Cumulus Linux extends Server Monitoring’s functionality to monitoring switches.

Once logged into Cumulus Linux, installing the server agent takes just a few minutes, as expected. Leveraging the documentation and installation guide allowed us to get up and running in minutes.

Since it’s SaaS, there is obviously no server deployment required, so all you have to do is to log in to your New Relic account and start looking at the performance data that is automagically pushed to your dashboard. Here’s Continue reading

Dell World: Driving Innovation in a New Era

I just returned from Dell World in Austin, Texas this past week. It’s one of my favorite events because not only does this week bring out hundreds if not thousands of new and innovative ideas, it also brings together thousands of people from around the world with different experiences and backgrounds.

Dell World booth - Cumulus NetworksThis event was like no other Dell event that I’ve attended. First, it actually rained in Austin. For two days straight! But more importantly (and as dramatic), as a private company, Dell had a very different vibe. Michael Dell repeatedly said that this was the first time he was having fun in a long time, and looking at the crowd, you could sense it! The general feel was that this new Dell could take on any new challenge and could do so swiftly. Always optimize for the long run and drive innovation to accelerate business.

It’s very evident that the software-defined world is changing. At DelCumulus Networks - Reza Malekzedah at Dell World 2014l World you could feel the buzz and witness each and every company’s evolution and observe how each product is helping evolve the face of the entire IT ecosystem. With all of the hype surrounding the software-defined market, I’m happy to have had Continue reading

Software-Defined Data Center Straight Talk with Tom Burns & JR Rivers

The data center is flush with change and it’s hard to know where to turn for advice. The myriad vendor positions on technology are confusing and seem self-motivated to lock you into their technology. Meanwhile, your data center is being flooded with more traffic every day.

Getting advice is hard so it’s not every day you get to talk with two of the world’s foremost thinkers in the data center world. With Tom Burns (Vice President and General Manager, Dell Networking and Enterprise Infrastructure) and JR Rivers (Co-founder/CEO of Cumulus Networks) joining me, I will get that special privilege on November 20 when I host a webinar to discuss open networking in the software-defined data center (SDDC).

Yeah, I know what you’re thinking. Another webinar with a bunch of marketing buzzwords. Well, if you know these guys you know that won’t be the case. I’ll moderate the discussion by teeing up a few questions and getting out of their way.

We’re planning to discuss a number of topics that will be sure to provoke some strategic thinking on your part. We’ll discuss:

  • Major data center challenges and how best to address them,
  • The current state of enterprise-ready SDDC technologies and Continue reading

OpenStack and Cumulus Linux: Two Great Tastes that Taste Great Together

OpenStack is a very popular open source technology stack used to build private and public cloud computing platforms. It powers clouds for thousands of companies like Yahoo!, Dreamhost, Rackspace, eBay, and many more.

Why drives its popularity? Being open source, it puts cloud builders in charge of their own destiny, whether they choose to work with a partner, or deploy it themselves. Because it is Linux based, it is highly amenable to automation, whether you’re building out your network or are running it in production. At build time, it’s great for provisioning, installing and configuring the physical resources. In production, it’s just as effective, since provisioning tenants, users, VMs, virtual networks and storage is done via self-service Web interfaces or automatable APIs. Finally, it’s always been designed to run well on commodity servers, avoiding reliance on proprietary vendor features.

Cumulus Linux fits naturally into an OpenStack cloud, because it shares a similar design and philosophy. Built on open source, Cumulus Linux is Linux, allowing common management, monitoring and configuration on both servers and switches. The same automation and provisioning tools that you commonly use for OpenStack servers you can also use unmodified on Cumulus Linux switches, giving a single Continue reading

Dude… cover me, I’m going in…

You know it needs to be done, it could be easy… or it could get messy, and you’re sure that the world will be a better place when you’re finished.

That’s the dilemma that some of our enterprise customers have when grappling with “the cloud.”

We’ve noticed a distinct trend among customers that grew up outside the “cloud era”; they’ve been trying to bolt “cloud” onto their legacy IT blueprint and it has been a struggle. They expected to realize operational and capital efficiencies that approximate high scale Internet businesses. Unfortunately, they are missing by a long shot.

At some point along the way, these customers realize that they need to be willing to drive structural change. They need to create a “cloud blueprint” for their applications and IT infrastructure. In some cases, this means a transition to public/hosted infrastructure; in other cases, it means building new private infrastructure based on cloud principles. In many cases, it’s a mixture of both.

When private cloud is part of the answer, we’ve consistently found design patterns built on infrastructure platforms like VMware vSphere and OpenStack and big data platforms such as Hortonworks.  Customers want to get these services operational quickly so they often stay with Continue reading

Accelerating Hadoop With Cumulus Linux

One of the questions I’ve encountered in talking to our customers has been “What environments are a good example of working on top of the Layer-3 Clos design?”  Most engineers are familiar with the classic Layer-2 based Core/Distribution/Access/Edge model for building a data center.  And while that has served us well in the older client-server north-south traffic flow approaches and in smaller deployments, modern distributed applications stress the approach to its breaking point.  Since L2 designs normally need to be built around pairs of devices, relying on individual platforms to carry 50% of your data center traffic can present a risk at scale.  On top of this you have to have a long list of protocols that can result in a brittle and operationally complex environment as you deploy 10′s of devices.

Hence the rise of the L3 Clos approach allowing for combining many small boxes, each carrying only a subset of your traffic, along with running industry standard protocols that have a long history of operational stability and troubleshooting ease.  And, while the approach can be applied to many different problems, building a practical implementation of a problem is the best way to show Continue reading

Open Networking for VMware vSphere: The Last Piece of the SDDC Puzzle?

VMware vSphere is one of the most prevalent virtualization platforms in the market today. Despite stiff competition from proprietary and open-source initiatives, VMware vSphere has continued to innovate and provide value to the enterprise.

Now that I’ve kept marketing/SEO happy, lets dive right in. Two of the latest initiatives from the VMware marketing engine are ‘“SDDC”’ (software-defined data center) and “Hyper-Converged.” Hype aside, these two concepts are fundamentally aligned with what hyper-scale operators have been doing for years. It boils right down to having generic hardware configured for complex and varied roles by layering different software.

At Cumulus Networks, we help businesses build cost-effective networks by leveraging “white box” switches together with a Linux® operating system. We feel this is the crucial missing piece to the overall SDDC vision.

sddc-cumulus networks

What is VMware’s SDDC and Hyper-converged Strategy All about Anyway?

First, let’s back up a little and talk a little history. VMware started as a hypervisor stack for abstracting compute resources from the underlying server (and before that, workstations… but I digress). They moved on to more advanced management of that newly disaggregated platform, with the rise of vCenter and everything that followed.

Then VMware turned their attention to the Continue reading

Nova-net is the networking component of Nova

There has been an aspiration to replace Nova-net with Neutron for about 4 years now, hasn’t happened yet.  The latest is that Neutron is being threatened with being demoted back into “Incubation”, due to promising to make itself production ready for each of the last 4 releases, and then totally failing to follow through.  All of the handful of production deployments of Neutron are in conjunction with Nicira/NSX-MH, which does all the heavy lifting.

The Neutron folks are optimistic that they will be production ready in Juno (the next release, Oct this year), but I’m betting on Kilo, the release early next year.

The post Nova-net is the networking component of Nova appeared first on Cumulus Networks Blog.