Data Center

Overview and Key Themes from NANOG 70

by asardell ‎06-09-2017 02:34 PM - edited ‎06-10-2017 08:56 AM (6,559 Views)

It’s great to be part of the evolving art of networking, and there’s no better place than NANOG. This week, network operators at NANOG 70 (Bellevue, WA) checked in with us continually for updates on the status of various solutions in our data center portfolio. I’ll talk about these and other questions in this post.   

 

First, the key themes addressed in the sessions and in the hallways were: 

 

  • Visibility and Analytics: there were several sessions on telemetry and measurement, and in private conversations we discussed how visibility is key to supporting our automation solutions
  • Security (DDoS): In the aftermath of the Wanna Cry worldwide cyberattack, savvy customers were keen to discuss best practices for keeping data center networks secure
  • Automation and Control: Consultations around automation continued to gain momentum in both the session presentations and in the many meetings we held with operators

Evolution of NANOG

 

NANOG over the years has evolved with the changing shape of the Internet. In its original charter and earliest meetings starting in 1994, it was dedicated exclusively to network service providers and the regular exchange of technical information to keep the Internet in sound shape.

 

It was always--and still is--about collaboration in the interest of Internet health and performance. In recent years, as edge data centers have changed the Internet’s geography, many more colocation providers, cloud service providers, and of course vendors supporting these trends, have joined hands in charting the way forward. So with this evolving demographic, nearly all of the general sessions were packed (Figure 1). 

 

gen session 1.jpg

Figure 1: NANOG Continues to Thrive with a Growing Audience (Source: NANOG)

 

Visibility and Analytics: Relevance to Applications 

 

Compared to the early days of NANOG, measurement and telemetry that once was only required at limited scale is now geared to the cloud, and requires collection and analysis at higher volume and at multiple layers.  

 

Accordingly, there were sessions on network telemetry from Princeton University and from Yahoo. And many other sessions talked about:

 

These lines of thought were encouraging to hear, as they led to further discussions around the Insight Architecture and Visibility Services in our entire SLX portfolio, particularly the SLX 9540, which was on display at the Beer-n-Gear session.  

 

Security (DDoS)

 

A key presentation in the general session was on DDoS trends and a call for action from longtime security architect Merike Kaeo. One of her key recommendations for best practices was to band together as a community and share attack information as long as it doesn’t impact privacy.

 

Thus, “attack use cases” such as SSH brute force attacks and DNS amplification attacks, as well as attempts at SPAM phishing, can be circulated among operators in order to quickly close the door on the damage before it becomes unwelcome national news. This viewpoint resonated with the audience.

 

In private meetings, we also discussed the primacy of security, and in some cases discussed the threat mitigation capabilities of Flow Optimizer, which has been further enhanced to support IP blacklisting and more platforms at higher scale in Release 2.0.

 

Service Automation and Control

 

Aspirations about improving automation and control were present in almost every conversation and in many general sessions as well. There was a lot of discussion around machine learning and its role in intent based networking and its many use cases.

 

There was general agreement that by asking the right questions of networks and ensuring that state is always mapped to policy, a paradigm shift can be made in data center infrastructure management.  

 

In part, this came up in many sessions on the setup and maintenance of Internet Exchanges and carrier networks, on the West Coast, Midwest, and in New Zealand.  Talking with some of these presenters after the sessions, we discussed how to help them thrive with hybrid cloud connectivity.

 

Even given all the exciting research in automation, architects and engineers from many service providers (even the most forward-looking) noted that they are still in the early days of going beyond the command line: mapping their configurations to rules and handcrafting the result into policies.  

 

To help bring this movement forward, we’ll be performing more public demonstrations of cross-domain automation with Workflow Composer (Figure 2) in future roadshows and events.

 

Server Based Automation Platform1.png

Figure 2: Server Based Automation with Workflow Composer

 

In Light of the Extreme Networks Acquisition

 

In the previous NANOG (February), customers were naturally curious about what was going to happen in the not-yet-acquired data center business. It was a pleasure to reassure these same customers that Extreme has publicly stated its goal of boosting its revenue significantly on the strength Brocade’s data center portfolio

 

Nothing says business as usual like a hard number. And in a recent online interview, Extreme CEO Ed Meyercord emphasized that our automation solutions and our hardware platforms (MLX, SLX, and VDX) will continue with their aggressive roadmap in force.

 

And the latter is very welcome news, because we received a lot of queries and excitement about the 100G and 25G density of the SLX 9540, and also about the recently available virtual SLX, which will be very useful for design and training purposes for our customers. 

 

In the end, customers were very glad to see the completeness of the portfolio (Figure 3) that is moving to Extreme.

 

SRA Porfolio1.png

 Figure 3: Switching, Routing and Analytics Portfolio

 

These offerings meld nicely into a pure play, end-to-end IP networking portfolio.

 

Call to Action

 

Contact your Brocade representative for information, demonstrations or evaluation copies: