Blog post

Networking Thanks

By Andrew Lerner | November 24, 2015 | 0 Comments

SD-WANNetworkingADC

US Thanksgiving is here in just a few days, so here is a very network-centric (and eclectic) set of things to be thankful for.

  1. Sniffers – I’ve never met a decent network engineer who didn’t love packet sniffers (i.e., ethereal/Wireshark/netmon). Nothing like pulling out a packet capture to show a server/NIC or firewall eating a packet. As my former colleague Todd Ferguson often said:  The Packets Don’t Lie.  For enterprise-class sniffing capability, look to network packet brokers.
  1. App Delivery Controllers (aka Load Balancers) – Getting called in the middle of the night is not fun.  Servers crash and networks get blamed.  However, App delivery controllers provide resiliency and scale to keep the app up, even when a server crashes.  Those of us who ran networks in the pre-load balancer era will better relate to this one…but worthy of being thankful for none-the-less.
  1. SDWAN –Years of incremental innovation have led to most enterprise WANs being complicated and brittle.  Further, there’s no rest for the weary, as BW grows at a 29% CAGR clip, and application architectures (cloud, micro-services, SaaS) are changing.  Not to mention that scaling a distributed complex system is exponentially harder than a simple one.  Never fear, SDWAN to the rescue.  Yes, the WAN is cool again.
  1. Netflow/IPFIX – because the only thing better than the actual data (see sniffers above) is fast access to metadata.
  1. Data Center Switching Fabrics – Who likes to do tedious box-by-box management on a bunch of disparate switches in the data center.  Maybe this was cool in 2005, but not really any more.  It’s all about single interface/console to manage the switches as a single entity or “fabric”.  Often, they include an API to hook into orchestration/CMPs and have automated switch insertion capabilities.  Saves time, saves money, reduces errors, wicked awesomeness prevails  (also see Gartner’s Technology Overview for Ethernet Switching Fabric).
  1. Internet/BGP Looking Glass – Ever troubleshoot an end-to-end problem across the Internet?  Good luck without having visibility into public looking glass sites.  There’s a decent listing of them here and here. Meanwhile, some of this functionality is being SaaS-ified (i.e., ThousandEyes), but “old-school” looking glasses are still cool in my book.
  1. Cloud-managed – Network engineers tend to enjoy managing traffic-passing infrastructure, but managing the management system isn’t nearly as rewarding.  It’s way easier to just plug it in, have it call the “mother ship”, and then do your configuration via a nice SaaS portal (just make sure to do a TCO).  Cisco Meraki, Aerohive, Aruba/HP, Glue Networks and many others support this emerging trend (see also Innovation Insight for Cloud Managed Networks).
  1. And of course, we would never have gotten this far without ping (or ping6). http://ftp.arl.mil/mike/ping.html

Leave a Comment