The future of a network administrator 2

[krajci]

It is becoming more obvious that as technology has moved back towards centralized processing power and away from PC based networks that decentralize the processing power. I have been a network administrator for NT and Novell based networks for the last four years. I realize that this change will take some time with many of the shops, but like the age of the almightly mainframe programmer our day will come too.

Where does everyone see the network administrators of today going tommorrow? [sig][/sig]

 

[THoey]

You see it as moving back to centralized power? I see it as going from centralized power (mainframes) to decentralized power (PC's linked together) to shared power (PC's linked to a server, each with parts of the whole task). And to link all of this togther, you need a network administrator.

Just my opinion. But then again, they don't let me out much...X-) [sig]<p>Terry M. Hoey<br><a href=mailto:th3856@txmail.sbc.com>th3856@txmail.sbc.com</a><br><a href= > </a><br>Ever notice that by the time that you realize that you ran a truncate script on the wrong instance, it is too late to stop it?[/sig]

 

[krajci]

We have been moving away from centralized systems for about ten years, but now with the true business advent of the internet you see more and more remote backup, ASP's, the slow death of Netware, the revival of UNIX. [sig][/sig]

 

[CitrixEngineer]

Network Admins are here to stay.

Business needs computer networks now more than ever before. Most businesses are absolutely reliant on services such as file-sharing and e-mail to the extent that a day without e-mail can cost a company millions. Only the network admins know how to keep it all together.

The way of the future is centralised software. Pay per use. ASPs. Microsoft loves ASps, because if people rent apps over the internet rather than buy physical media, they can't copy it. Copy it = piracy.

ASPs depend on huge teams of sysadmins to keep the banks of servers running. QED.

More and more companies are also beginning to enjoy the lower TCO of thin client computing. While this makes MSCEs and junior tec support tremble in their boots, it's just another job for sysadmin. [sig][/sig]

 

[kjonnnnn]

They WEEW moving away from central power as far as Mainframe, But a net admin is needed more that ever now to hook all these computers together. Even the most savvy 'lay person computer geek,' doesnt know a whole lot about networking.... [sig][/sig]

 

[radiohead]

The way I see things heading is specialization to the extent where you will have server admins, storage admins, desktop admins, database admins, network admins all in their own little worlds. The few who have an understanding of it all will probably end up as consultants, talking to each of the individual teams who are blaming each other for the problems which occur when people don't understand the interconnectivity.
A bit like programmers and admins now;-)

 

[CitrixEngineer]

(edit). Warning: This is very long. I really got carried away and enjoyed writing this piece. I hope it is enjoyed by those who read it. (end edit)


OK, Radiohead (excuse cheap joke!), I think, if you look around the average IT department, that what you describe is how things currently are.

Things are gradually moving towards the Citrix way, and so the scenario you describe will soon become a thing of the past. No longer will someone be able to wave a piece of paper labelled &quot;MSCE&quot; or suchlike, and get a job in IT on that alone.

I hope I will not be severely chastised for shameless plugging, but here is my vision;

With the Citrix solution, the need for desktop support will decrease substantially. Systems administrators and network administrators roles will blur so that the distinction will be minimal. Database administrators will become more specialised in their roles, it is true, but they will be able to concentrate on the task in hand - solving the complex problems involved with large relational databases - rather than wondering whether there is a machine or network problem. The same goes for software engineers.

End users will benefit greatly too. Instead of having to pay $400 for the latest Office suite, it can be rented over the internet. With the advent of 0800 ISPs, on-line time costs nothing, and the only costs are incurred as you use the software. If you use it a lot, you pay more. Just like gas. Plus any updates are done for you.

Home PC crashes will become a thing of the past, too. When all software is run from the internet, not the local machine, there will be little but hardware failure to bring a PC to its knees.

Software companies will benefit, too: If the software is run over the web, stored on the company's servers, the risk of piracy is hugely reduced. The benefits from this can then be passed onto the consumer.

Mobile communication will be revolutionised: Imagine checking your e-mail or browsing the web - or working on an important office project with live data - from a beach, via your hand-held device, your mobile phone or one of the &quot;next generation&quot; communicators, which combines the two.

This is not some Utopian dream. This is what can be done NOW with Citrix technology.

I'm not a salesman or a company representative, just someone who works with this technology and sees a real future for it, professionals who adopt it, and its customers.

If you want to get ahead in IT, learn this technology. There are companies in London offering £50-100k for good Citrix admins. [sig][/sig]

 

[radiohead]

I agree this is a very good way to go, but as someone who recommended a citrix solution for managing approx. 100 remote sites, I know others still don't think this way.
We ended up implementing a server & new workstations at each site - I don't really want to think about the costs.

We'll have to see, but my gut feeling is server admins will be somewhat like ops and the directory service will be all. How that pans out (stand alone machines or thin clients) is anyones guess (which is what this discussion is about).

Anyway, good luck.

 

[krajci]

This is actually exactly what I was talking about.

I have managed a Citrix Metframe server with approximately 50 thin client (that run at about half of the cost of a PC in most cases) and it had some distinct advantages as thin client a normally dead or alive, not much grey there. Some of the disadvantages, of course, were in the case of a line issue (which we had on occasion with our carrier, they couldn't even type a memo.

The administration of a Citrix-based network is cake, but the cost can be very restrictive. In addition to that, if you exceed 50-60 workstations on a Metaframe server doing 'common' office activities (MS Office, a little database work, etc.) you normally need to add another server to your farm. Also, you do noy go with a light server to run Metaframe.

With all these advantages and disadvantages of central processing versus decentralized processing it still seems that I see more movement back to the days of the central mainframe and thin clients, its just now a Citrix-type solution (or the ASP increasing marketplace) instead.

This communication is very interesting though seeing all the varying viewpoints here. [sig][/sig]

 

[CitrixEngineer]

Radiohead

I don't understand why you had to install a server at each of your 100 sites if you used MetaFrame! The whole point is that it works over high-latency connections.

The theory (and I've put this to practice many, many times) is that you have a bank of centrally located servers, and remote sites hook up to them. If the servers are local to each other, you'll get better replication propagation and overall control.

Yes, the initial outlay is very high. That is usually where the expenditure ends, and the savings begin, in my experience:

As you say, krajci, the administration is &quot;cake&quot;. When it comes to upgrade time, you upgrade software once. On the server - not for each client. The client devices should never need to be upgraded. The same client *could* be connecting to a Citrix server inthe year 2010.

You don't have to buy expensive new kit for the clients either:

I've got a demo 286 that I like to show to my disbelieving colleagues. It's always a pleasure to see their faces when I prove that it can *run* Office 2000 at least as fast as their brand-new desktop.

50-60 users at one of my sites still have Windows 3.11 on their 486s connecting to a Novell 4.11 network. If they need up-to-date apps, they connect to Citrix published applications, otherwise they're happy with what they've got - why change and introduce new problems?

As for the cost to them, well, a Brand new quad processor server with MetaFrame cost around £30k. If we'd bought *decent* PCs, say £800 a time, we'd have had to double the budget - just for initial costs, never mind the ongoing support.




[sig][/sig]

 

[krajci]

I did not use a server per site, this must have been a misunderstanding/misstatement, but one server at a centralized location ran all the sites.

I can see the arguement for Metaframe, and actually have tried to get it in my current shop for some internal use and for porting sessions over the internet. [sig][/sig]