The Fat Pipe - All Packet Pushers Pods

Today on the Packet Pushers’ Priority Queue, we catch up with GNS3. What’s GNS3? It is a free and open source network simulator, sponsored by SolarWinds, that’s incredibly useful if you’re trying to lab and learn without all that pesky physical equipment.

If you’ve not heard of GNS3, you’re in for a treat. If you’ve heard of GNS3, but haven’t used it in a few years, listen in to discover just how far the platform has come.

Joining the Priority Queue from GNS3 is Instructor David Bombal, Principal Developer Jeremy Grossmann, and Developer Julien Duponchelle.

We discuss GNS3’s origins as a certification practice tool and its open source status and relationship with SolarWinds. We also discuss the capabilities of GNS3, the routing, switching, and middlebox technologies supported (not just Cisco), and how GNS3 compares to other network simulation platforms.

Special Offer:

SolarWinds has a special offer for Packet Pushers listeners. Get the SolarWinds Standard Toolset for free! The Toolset includes network monitoring, performance monitoring, and remote management. Just visit tinyurl.com/packetpushers to get your free license key.

Links:

GNS3 Academy Courses

GNS3 on Twitter

David Bombal on Twitter

Julien Duponchelle on Twitter

Julien Duponchelle on GitHub

Recorded live in New York, network engineers talk candidly about their process of deciding to buy and deploy a Viptela SD-WAN solution. This live, face-to-face discussion, sponsored by Viptela, gave us one of best podcasts of 2016. We were joined by:

* Eric Murray and Kyle Alexander from Kindred Health Group

* Gant Estes and Jarin Dykstra from 77 Energy

* Nick Phelps from FreedomPay

Here’s a few key points that grabbed my attention. We kick off talking about features that mattered most to our guests. Getting “free” from the telcos and carriers was a fun concept, and being able to negotiate from a position of strength puts smiles on peoples’ faces. The ability to identify and isolate traffic flows as “micro-segmentation” and the use of traffic engineering for both SLA and encryption really do solve business problems.

For Kindred Healthcare, IT is a shared service and locations may or not be fully under their control. SD-WAN means moving the trust boundary closer to the network edge (not normally possible for a WAN).  Importantly, they were able to implement traffic steering without implementing MPLS and to use a centralized point of control to reduce configuration time.

For 77 Energy, a key feature is that just two people can operate their 500 WAN locations, and they don’t expect that number to increase as they roll out more sites during planned expansion. As a recent spin-out from a larger company, the ability to scale up/down their exploration sites on short notice has been a success.

FreedomPay has unique requirements as a payment provider. Encryption and device security are critical, but a surprise was the Viptela SEN feature, which let them build arbitrary network topologies in partial or full mesh depending on the application.

One more thing that I haven’t heard previously is that operation of Viptela devices using a CLI can be reassuring, especially when configuring IP routing protocols for integration with the legacy WAN.

A final standout was hybrid cloud connectivity, where it turns out that AWS and Azure don’t have enough features for connecting your SD-WAN. Both 77 Energy and FreedomPay built unique solutions that deliver much more flexible outcomes while still connecting the WAN directly to public cloud.

Thanks to Viptela for sponsoring this show and making a live podcast possible in New York.

Wireless. It’s what’s carrying your data. And today, Cambium Networks sponsors the Packet Pushers for a discussion about WiFi in your enterprise as well as a new solution they recently announced that offers 14×14 MIMO. Enough with your piddly 3×3, 4×4, or 8×8. Cambium’s got 14×14.

Joining us for today’s conversation from Cambium are Puneet Batta, director of engineering for WiFi; Chetan Hebbalae, senior director, product management; and Trevor Miranda, software architect.

We discuss Cambium’s products, including cnPilot, cnMaestro, and its newest AP cnMedusa (that’s the 14×14 MIMO). We also talk about the ins and outs of wireless broadband and recent developments in enterprise WiFi.

Cambium will offer a 30% discount for listeners who register at www.cambiumnetworks.com/packetpushers and evaluate their cnPilot E400 and E500 access points.

To learn more about Cambium, check out our previous podcast, PQ Show 74: Cambium Networks – Affordable PTP Wireless (Sponsored).

Containers are a hot topic. While most people are focused on microservices, the Packet Pushers are thinking about how containers connect to a network.

What if you have containers in your private cloud and you want to use the same provisioning and monitoring tools in the public cloud? How do you set and monitor policies and performance? What does network operations look like when you don’t have a switch to configure?

Nuage Networks, maker of the Virtualized Services Platform and sponsor of today’s show, is here to address these and other questions.

Joining us are Harmeet Sahni, Director of Product Management at Nuage Networks; and Rajat Chopra, Senior Software Engineer at Red Hat.

We discuss container networking requirements for a variety of scenarios, including connecting containers on the same host, host-to-host networking, and host-to-outside world, and dig into the use of orchestration tools like Kubernetes.

We also explore the role of overlays in a container environment, look at service insertion, and more.

To get more details about Nuage Networks and containers, you can download this solutions brief from Nuage.

The Open Compute Project (OCP) describes itself as “…reimagining hardware, making it more efficient, flexible, and scalable. Join our global community of technology leaders working together to break open the black box of proprietary IT infrastructure to achieve greater choice, customization, and cost savings.”

Started by mostly large organizations with massive buying power (including Facebook), it might be easy to dismiss OCP as a project for other folks. But the OCP is starting to affect the market in many areas, including networking.

To catch us up on what the OCP Networking Project is up to, co-chair Carlos Cardenas sits down with the Packet Pushers. Carlos has written some of the most outspoken articles on PacketPushers.net, having zero trouble speaking his mind, so this will be an interesting episode.

We’ll talk about why networking was added to OCP, its efforts around designs for data center leaf and spine switches, what’s come out of the project to date, and what efforts are being targeted at the campus, including access switches and wireless APs.

Show Links:

Open Compute Wiki On Networking

OCP Networking Specs And Designs

Open Compute Mailing List

Carlos Cardenas On Twitter

Carlos’s Posts On Packet Pushers

In the beginning, there was SNMP with its MIBs and OIDs. And then there was rrdtool, MRTG, and cacti, graphing our little OID values for us in a symphony of contextless images. They looked impressive, but tended not to tell us much.

Network monitoring has evolved, and network monitoring tools along with it. Our sponsor today, Paessler, is here to discuss good and bad monitoring practices, and just how they’ve kept up with the times.

Joining Greg and Ethan, from Paessler is Greg Ross, Senior Systems Engineer. We’ll get an update on Paessler’s PRTG Network Monitor, talk about how to do monitoring right in complex networks, and how to get actionable information from disparate data points.

We’ll also talk about how to identify signs of trouble before something breaks, and get tips for working in cloud environments.

Special Offer

And Paessler has a special offer for Packet Pushers listeners: get a free network consultation when you download and install a free trial of PRTG Network Monitor. Get details and sign up here!

You can also follow Paessler on Twitter, LinkedIn, and Facebook.

Today’s show was recorded live at VMware’s Future:Net conference 2016 in Las Vegas. Greg Ferro and Ethan Banks talk with Bruce Davie, CTO, Networking at VMware; and Simon Crosby, CTO and co-founder of Bromium, about the future of security in the context of networking.

The discussion looks at security both on and between end points, and how much security can be done in the network. Do middleboxes still have a roll to play? Are there other ways to get end-to-end security? Is the endpoint the best place to protect the organization?

We also look at advances in attacker techniques, why it’s nearly impossible for the enterprise to defend its boundaries, and how microsegmentation affects the security picture.

Join us for a lively discussion with two technology luminaries about where security is going and the network’s role.

5G networks are still a ways out from deployment, but the technology will dramatically boost the capability of mobile networks by providing 1Gbps of bandwidth and lower latency, and by taking advantage of a wider spectrum range.

Besides mobility and wireless, 5G will likely have a significant impact on industrial networks, IoT, and cloud computing.

On today’s sponsored Priority Queue, we talk with Intel about 5G technology, how it works, business use cases, and its potential impact on the enterprise. We also discuss Intel’s efforts to bring 5G-enabled products to market. Our guest is Caroline Chan, Director, Wireless Access Strategy and Technology, at Intel’s Network Platform Group.

This podcast was recorded live at the Intel Developer Forum 2016 in San Francisco.

Links:

Network Transformation Is Laying the Groundwork for 5G – Intel

A Look Ahead…The Network Of 2020 and Beyond – Intel

5G Technology Overview – Intel

A 5G Network is Possible Now on Intel Architecture – Intel

Paving the Road to 5G Mobile Services – Intel

How 5G Will Power the Future Internet of Things – Intel IQ

Powering 5G Network Infrastructure with Intel Technologies – Intel

This Priority Queue episode was recorded live at the IETF 96 meeting in Berlin, Germany in July 2016. The discussion focuses on the IETF working group I2RS (Interface to the Routing System).

According the charter, I2RS “facilitates real-time or event driven interaction with the routing

system through a collection of protocol-based control or management interfaces. These allow information, policies, and operational parameters to be injected into and retrieved (as read or by notification) from the routing system while retaining data consistency and coherency across the routers and routing infrastructure, and among multiple interactions with the routing system.”

Greg is joined by two guests to explore how I2RS works and its benefits. First is Sue Hares, a co-chair of the I2RS working group. She’s also a consultant at Hickory Hill Consulting and an expert on routing.

Also joining the conversation is Russ White, a network architect and frequent Packet Pushers guest.

The Priority Queue dives deep today, with a discussion of using EVPN as an enterprise data center interconnect. This is not a feel-good, fluffy podcast about how nice EVPN is and the fun things we someday hope to be able to do with it. Rather, this show gets right into the weeds on EVPN with the specific use-case of DCI.

Join Ethan Banks as he interviews double JNCIE Clay Haynes on setting up BGP adjacencies, creating EVPN address families, interpreting BGP output containing EVPN NLRIs, handling VTEPs, walking a packet through end to end, and useful commands to know.

More about EVPN?

* Show 196 – EVPN Introduction & Use-Cases With Russ White + Jeff Tantsura

* Show 233 – Cisco Nexus Using BGP As A VXLAN Control Plane – Sponsored

* EVPN: Intro To Next Gen L2VPN

Other resources.

* MPLS in the SDN Era

* Rick Mur’s Blog – EVPN (RFC7432) Explained and EVPN Configuration

* Dan Houtz – Configuring EVPN on Juniper QFX

Want to follow Clay?

* Twitter: @dark_15

* LinkedIn: dark1587

* Blog: http://alostrealist.com

Show Notes:

Section 1 – What is EVPN?

VPN – not IPSEC, but a virtual private network nonetheless.

Provides L2 or integrated L2+L3 VPN.

Implemented in the control plane using multiprotocol BGP.

Implemented in the data plane with your choice of…

MPLS

Overlay such as VXLAN

Provider Backbone Bridging (PBB)

Section 2 – What is EVPN good for in the real world?

In a nutshell, it carries entire Ethernet frames across an L3 infrastructure.

When using MP-BGP as a control plane, MAC addresses are advertised between BGP peers.

A major use case for this is DCI.

Stretching L2 between data centers is a design problem that needs special handling to implement safely.

Extending failure domain of STP.

Sub-optimal FHRP = traffic trombone.

EVPN handles some of these stretched L2 issues.

What vendors support EVPN on what platforms?

Cisco

Juniper

Other

Section 3 – Let’s implement EVPN as an enterprise DCI.

First, some terms.

Ethernet segment

Ethernet tag

Ethernet segment identifier

EVPN Instance (EVI)

Then, some components

EVPN overlay format of NVO.

VXLAN tunnels & VTEPs

MP-BGP peers sharing EVPN data

MACs

VNIDs

Packet walk time! I’ve got these two endpoints on the same stretched L2 segment that need to talk to each other, separated by L3 infrastructure…

Lessons learned.

Architecture traps to avoid.

Redundancy concerns.

Troubleshooting tips.

Reading the BGP table.

Useful commands to know.