I am sitting here in my green chair at home listening to 'Desperado' by the Eagles and reading my friend Greg Ness' recent piece on Infrastructure 2.0. I've done a few panel discussions with Greg in the past, including a really fun one at the FiRE conference earlier this year, and yet in a past life at the same time.
Greg makes some interesting points though about how infrastructure needs to change, and he used the example of an industrial-era factory. In manufacturing we saw a major shift in the the past decades from traditional build-warehouse-sell-ship models to just-in-time manufacturing. I remember my father studying for weeks for his APICs tests as this major shift was rolled throughout major manufacturing shops and the new processes were put in place. (Coincidentally, Dad ended up as the first CFPIM in AT&T/Western Electric/Lucent back then, although I really didn't have any appreciation for that at the time but learned that I must avoid a career in manufacturing planning at all costs at my dad's knee)
IT is at a similar precipice- will we move from build it they will come 'Field of Dreams' models to a Just-in-Time IT architecture? What will the impact be on the infrastructures we have built? What will the impact be on the role of the network?
My fear, that may be unfounded - but worth discussing, is that the next wave of IT value will not be delivered by the network but instead by software and virtualization platforms. Systems vendors, such as HP, are moving to reduce the network to dumb plumbing. Budgets seem to be shifting to projects that directly influence the business top/bottom-line, and away from 'plain ol' infrastructure'.
As someone who has spent the majority of my professional career as a 'network guy' I really don't like this- but if I accept that this may actually be happening I have to ask the question -- If the pendulum is shifting to software and virtualization projects, if the network is getting taken for granted or worse being seen as a 'blocker' to the deployment of these newer models, what should network professionals do and what should networking vendors build to be a part of this architectural transition?
dg
Douglas Gourlay is the vice president of marketing at Arista Networks - a leading developer of 10Gb Ethernet switching platforms. In this role Gourlay is responsible for the global marketing and product management for Arista.
Prior to joining Arista Networks Gourlay was the vice president of Cisco’s Data Center Solutions Group, where he defined and executed Cisco’s global marketing strategy for data center, virtualization, and cloud computing. This included the Nexus and Catalyst data center switches, application and server load-balancing, storage networking, blade switching, and wide-area application services product families. Under his leadership Cisco’s data center segment grew from a nascent business to over $5B in annual revenue.
Since 1998 Gourlay has led and contributed to numerous hardware, software, and systems architecture developments across Cisco. He has served as senior director of product management for the Nexus Family of data center switches, director of product management for the Catalyst 6500 Series of LAN switches, and led product management for Cisco’s Application Delivery product family. Gourlay has filed or holds more than 20 patents in networking technologies.
Prior to his work at Cisco, Gourlay was an industry consultant and served as a US Army Infantry Officer. Gourlay is an avid pilot and can often be found tinkering on his Cirrus at Palo Alto Airport.
The pendulum also swings back the other direction...
Dare I point out the obvious that this has happened before and it will happen again?
If it's a pendulum, unless one redefines the laws of physics or the bob in your example simply stops, it swings back the other direction, too.
When I gave the first presentation of the 4 Horsemen of the Virtualization Security Apocalypse almost 2 years ago, people gasped at the notion that the network access layer would be subsumed by virtualization platforms.
I followed that up with the reflection that within 5 years with IOV (before the unified fabric/unified compute moniker) virtualization and security folks ought to run out and get their CCIE because it was going to come full course again and the "network" would evolve to include abstraction at levels not seen before...including delivering apps directly.
So that's my opinion. I've been pretty lucky with my guesses thus far... ;)
/Hoff
That's why I love you Hoff (SQUIRREL!!!!)
Hoff, I agree- have always seen things come full circle. Some people comment that the data center of today is simply a re-consolidation into the mainframe world of the past, just this time with different CPU architectures and open-standard interconnects.
What do you think a network engineer/operator should do? Should they learn about virtualization platforms and the impact those will have on the network? What can they do to ensure that in the interim disruptive period they keep the network relevant, or is it a non-issue?
Follow me on Twitter or on loopback0, my tech blog
The Return Swing isn't a Guarantee this Time
I htink the question relates to who delivers on the promise of automation and policy control first. If the virtualization and systems vendors produce the cleanest path they could render the network irrelevant. In the HP vs Cisco war, the winner will be the player who solves these issues first. And the winner will have a major impact on the shift in IT careers.
AT this point it seems clear that until the network effects of virtualization are addressed virtualization will continue to spread tactically and with "contained" benefits. That may help to preserve a dying status quo a bit... but eventually the box will open and there will be substantial career shifts away from silo specializations and manual taskmaster roles.
My twitter account is www.twitter.com/archimedius
G
Dumb plumbing
With everything being sent to Asia and wages being globalised downwards no one will have the experience or the wish to study in this country to implement anything intelligent
I hate to say this, being a
I hate to say this, being a network guy myself, but the network is already dumb plumbing. The network itself isn't worth much without the devices that communicate across it. That being said, the converse is true as well. I don't care how many servers can be virtualized on a single hardware platform or what kind of applications it can run, it still has to communicate with other devices.
The performance of virtualized devices relies, to a significant degree, on the ability of the network move traffic effectively and efficiently. That is where the virtue of the network lies. The scope of the role of the 'network guy' might change, but the network will continue to be an integral part of a company's ability to deliver services to its customers.
what should the upcoming
what should the upcoming networking guys do?
what we've always done
Shift happens. And it happens faster for us in IT than for most. I don't know that this is any different for us in networking than it is for anyone else.
Given that--I think networking professionals + vendors have to learn to speak the language of applications and business impacts just like everyone else does. And be able to tie every networking product, service, offering, whatever to those things in demonstrable way. We've been crap at it, for the most part. That's gotta change.
Similar to server evolution?
As mid-level servers became commoditized, the skillset required to maintain and fix them was lowered. With virtualization, the complexity of the overall server/storage environment increased again, and introduced new networking complexities. I think the same thing's happening in networks as a whole. Network virtualization is gathering pace. So I think there'll always be a role for guys who know what's going on under the hood, particularly as virtualized networking spreads over the physical infrastructure. Plumbers are always in demand because most people just don't get plumbing :)
The next step may be plumbing for nonplumbers
Given the rise of system automation.
Post new comment