Industry execs: Network admins an endangered species

Jackace

First let me say, I know this topic isn't directly related to the CCIE certification, but the certification is mentioned in the article. (Link at bottom of post)

Second, I was hoping to get input from those of you with more experience than I on this topic.

It seems everywhere you turn someone is saying that network admin/engineering is going away. I remember when they said the same thing about system admin/engineering, but we never saw it happen. If anything we have more of them today then we did before because the technologies are much more diverse. I tend to think the same thing will happen in networking. Personally I think articles like this one are put out by sales and marketing people to try and bring attention to their new product. I don't doubt that a network admin/engineering job is going to be different in 5-20 years, but I'm not so sure the jobs are going to go away like these people are predicting.

So what do you all see as the future of our profession?

Industry execs: Network admins an endangered species ? The Register

YFZblu

In the end, that's a sales pitch - as far as the main point is concerned, as IT personnel we're considered an expensive overhead cost; business enablers at best. This will always be a sticky thing with business-side leadership.

Part of our job is to adapt. Do I think automanaged networking / virtual networking is going to wipe out the industry overnight, leaving networkers stranded? Absolutely not - The idea of getting rid of the people who know how the bits are moved through the organization and beyond is laughable to me.

gorebrush

Yeah we've had Network Automation tools here for a while. We still have the same number of Network Engineers too, because as a company we haven't really capitalised on the capabilities of it.

N2IT

I think IT in general has scaled back in some areas and increased in others. It's just the nature of business and information technology.

wintermute000

The automation they talk about in that spiel is the CCNA level automation (placing hosts into VLANs etc.)
It doesn't affect CCNP/IE level stuff.

YFZblu

wintermute000 wrote: »

The automation they talk about in that spiel is the CCNA level automation (placing hosts into VLANs etc.)
It doesn't affect CCNP/IE level stuff.

Did you read the article? The Juniper executive specifically calls out CCIE's as being affected by this.

Jackace

YFZblu wrote: »

Did you read the article? The Juniper executive specifically calls out CCIE's as being affected by this.

This is why I posted the article here. I personally don't believe all the hype and I don't think it is going to happen anywhere near as quickly as these articles seem to indicate, but you never know I could be totally wrong. That is why I came here to ask the question.

JoshyJ

I always get a bit down reading these type of articles.

It will be interesting knowing Cisco's view. Anyone know if Cisco has ever had a word in regards to this?

NetworkVeteran

JoshyJ wrote: »

It will be interesting knowing Cisco's view.

Cisco is all for this. If you worked for a network equipment manufacturer, and you could pitch customers to pay an extra ten million in equipment because it will save them five million per year in labor expenses, of course you would. Now, the future is not all rosey for them, either. New technologies threaten to reduce the need for very intelligent network devices. And remember nothing happens overnight. As long as you continue to adapt and stay abreast of new technology, I wouldn't worry much.

sides14

Just a rehash of articles that have all been written in the past. Of course personnel overhead costs are expensive, but realistically, it is cheaper than paying a vendor for their services and the corresponding downtime of equipment. Companies have been selling the idea of an "intelligent network" for years. Ultimately, a machine cannot fix a damaged cable or replace a faulty component. Eventually maybe, but not anytime soon.

Ahriakin

Articles like this always make me laugh. I've been hearing this stuff since I started in IT...remember how Windows 95 was going to put support staff out of a job? Some idiot exec makes light with hyperbolae (usually connected in some way to their own product) and a journo who either knows no better or is simply lacking for anything else interesting to publish runs with it. Network management tools and new processes/software are just about matching the increases in network complexity, the abilities required to design and support those networks is coming from a finite resource - intelligent and talented engineers.
Also at the base level their cost assumptions are way off. Our team for example has designed/installed enough hardware in the last 2 years to pay our salaries for the next 20, so stating that the human component is the primary expense is just laughable.

colemic

'As an example, he said that HP's technology services people have told him that to set up Exchange in a corporate campus takes a minimum of two weeks."That's the time, the human middleware factor," he said. "That's very concrete." Using HP's network-configuration automation, "they have seen it go to five minutes, and deprovisioning in seven minutes."

He's comparing apples and oranges here. I'll eat my hat if someone (or a software program) can install and configure Exchange correctly in 5 minutes. Unless he's talking about cabling/physical installation... which can't be the case, it doesn't take 2 weeks to install a server in a rack. Unless you have multiple other issues/complications that negate the usefulness of automation.

It's a shill pitch. Netadmins aren't going anywhere. If you are in a compliance-regulaed industry (DoD, financial sector, etc.) especially, your examiners/auditors/accreditation officials aren't going to find it funny if you try to convince them a computer can evaluate potentially subjective settings and make decisions. Not only that but, even if they do get their whiz-bang stuff working, if they aren't working on the compliance/evalaution side, no one will use it. No one will want to automate something that they can't ensure has been reviewed and is in compliance with ever-changing policies and standards.

Just my .02

sieff

This article seems very pro-HP and its' virtual automation products. IT is a cost center, but when things go wrong the value of having a competent IT staff is unparalleled. Most companies don't need a CCIE on staff to manage infrastructure once it's deployed. Only the large enterprises are willing to pay the higher salaries to retain top talent. Companies are already seeing a weakness in sales in network infrastructure components like routers and switches due to cloud services, but that doesn't mean automation products will be the end of the network admin. It's possible for companies to manage more with less personnel now, which we're seeing that in the current economy. As technology makes more progress I think there will be need for more skilled and well-rounded network admins. So your Exchange guy will need to know how to build a VM. The route & switch guy will need to know how to tune wireless or make adjustments to security appliances, etc...

vinbuck

This is an SDN pitch plain and simple. The reality is that SDN over time will redefine the way we look at network provisioning and operations, but like everything else will happen over a long period of time. Also, this will merely shift the roles of engineers and admins slightly to encompass this new technology.

Overall, I think it will be a good thing as it will get us out of the edge game via automation and back into the core/distribution world - which I think is more fun anyway

MiikeB

There are plenty of companies out there now using 10 year old technology, I see no reason to think in 10 years there won't be companies using what is now considered current technology.

lrb

I like SDNs and i think they have some very good use cases in production networks. However, don't be fooled into thinking we are all going to be replaced with software developers and automated tools any time soon - SDN isn't going to be the silver bullet that these guys seem to think it will be!

Jackace

lrb wrote: »

I like SDNs and i think they have some very good use cases in production networks. However, don't be fooled into thinking we are all going to be replaced with software developers and automated tools any time soon - SDN isn't going to be the silver bullet that these guys seem to think it will be!

I agree. SDN looks great, but it is going to be years before they have it perfected and even then you still want someone around that knows what all this automation and centralization is doing to fix things when (not if) it fails. For me what this stuff means is eventually there will be a smaller need for true network experts, but the need is not going to go away because of these tools. These tools will just make the true experts more efficient.

NetworkVeteran

Jackace wrote: »

I agree. SDN looks great, but it is going to be years before they have it perfected

As soon as a company discovers SDN is cheaper / easier-to-manage than their non-SDN solutions, you can expect them to begin dabbling with it. Perfection isn't necessary. I am already playing with SDN equipment and my employer is certainly not the only one doing so. Is it a silver bullet curing all networking wrongs? Of course no. Does it have dramatic transformative potential? Definitely. I will be excited to see what high-end networks look like five years from now.

Jackace

NetworkVeteran wrote: »

As soon as a company discovers SDN is cheaper / easier-to-manage than their non-SDN solutions, you can expect them to begin dabbling with it. Perfection isn't necessary. I am already playing with SDN equipment and my employer is certainly not the only one doing so. Is it a silver bullet curing all networking wrongs? Of course no. Does it have dramatic transformative potential? Definitely. I will be excited to see what high-end networks look like five years from now.

I totally agree that companies are dabbling with SDN as they should be, but from what I have seen we are still a good many years away from standardization with SDN. Then after standardization we have to get all the different vendors into compliance with those standards.

NetworkVeteran

Jackace wrote: »

we are still a good many years away from standardization with SDN. Then after standardization we have to get all the different vendors into compliance with those standards.

That happened years ago, although it seems to have really coalesced about a year or so ago. There are already compliance test tools on the market. The IEEE hasn't spoken, but our industry has a record of not waiting for them.

The primary SDN standard is already supported or soon to be supported according to Wikipedia by Big Switch, Brocade, Arista, Cisco, Force10, Extreme, IBM, Juniper, Larch, HP, NEC, and MikroTik.. and I know other big players not listed there.

In April 2012, "Google essentially has remade a major part of its massive internal network,
providing the company a bonanza in savings and efficiency" --- Wired Magazine

But many will stay somewhat mum until their big roll-outs.

Jackace

Sorry I did not mean to say there was no standard developed. I know Openflow is pretty much the current standard, but some of the blogs I have read say other players are still working on alternative standards or changes to the current standard. From what I have read and listened to the big problem with Openflow right now is the lack of APIs or the lack of good APIs.

Also the only big player we have really heard much about is Google, and they built their own hardware to support their network. I'm sure others are doing something similar to what Google did, and I agree with you that we won't hear much if anything until the big roll out happens, but I just don't see all that many places changing over in the next 2-3 years.

pert

I think its inevitable that networking gets easier and easier in the NA level realm. I think this will really, really impact newer engineers and people trying to get in the field. Even now, people with CCNA level knowledge have less and less reason to exist in a corporate environment. I think you'll see the number of people in the field go down, the competition get fiercer, and the top % will continue getting more and more pay. I already feel like there is no easy / realistic path to go from fresh out of school and get on a track toward being a senior network engineer. Its hard to find the rails for your train. Compared to other fields where the path from new to master is well established its night and day.