The Quiet Before the Next IT Revolution 145
snydeq writes: Now that the technologies behind our servers and networks have stabilized, IT can look forward to a different kind of constant change, writes Paul Venezia. "In IT, we are actually seeing a bit of stasis. I don't mean that the IT world isn't moving at the speed of light — it is — but the technologies we use in our corporate data centers have progressed to the point where we can leave them be for the foreseeable future without worry that they will cause blocking problems in other areas of the infrastructure. What all this means for IT is not that we can finally sit back and take a break after decades of turbulence, but that we can now focus less on the foundational elements of IT and more on the refinements. ... In essence, we have finally built the transcontinental railroad, and now we can use it to completely transform our Wild West."
Re: (Score:3)
Now - just because one company goes belly-up doesn't mean that another can't take over and be successful.
What you have is not by far a successful IT platform yet, you have the foundation. What is limiting is the ISPs and their customer agreements that effective limits the users to being consumers of bandwidth and services. When the ISPs realizes that their models with bandwidth throttling and agreements prohibiting customers to set up services at home slows down development of new companies and services the
Re:Horseshit (Score:5, Insightful)
I envy your optimism and agree that ISPs are the problem, but I don't see how new companies and services will force change upon ISPs.
New ISPs? Not in the state-sanctioned monopolist USA.
Loss of customers? See above.
The ISP and backbone provider bridge trolls sleep soundly, knowing that no one has the money or statutory permission to build competing bridges.
Only the FCC and Congress could do that, and the oligarchs are quite happy with the current bridge trolls.
Re: (Score:2)
That's only in the US. Here in the EU we solved that problem long ago, paving the way for the development of new companies and services. You will be left behind if you keep letting the market decide these things.
Re: (Score:2)
Re: (Score:2, Interesting)
I know that was sarcasm, but for the sarcasm impaired (or the ignorant), I recommend reading Greenspan's testimony to a congressional oversight committee in 2008 where he was forced to admit that the objectivist-based idiocy about free markets and rationality always winning out that underpinned the Reagan Revolution and subsequent de-regulation and freeing of the "free market" does not work in the real world.
Amazing, to see someone who gazed admiringly on Ayn Rand as he sat at her feet forced to admit his e
Re: (Score:2)
And if he had known his history he would have seen the stupidity of his ways. The history of financial crashes of the 1800's, before financial regulation was wide spread or even conceived of in some cases, is compelling. See en.wikipedia.org/wiki/List_of_banking_crises
Real world data refutes the market deregulation.
Stupid (Score:2)
No one argues for unregulated markets.
Reasonable regulation, built on experience, is all that people ask for and all that is needed.
Forcing companies to provide mortgages to people who are patently unqualified is an example of unreasonable regulations that resulted in untold devastation to the economy.
Now, the Feds are going around telling banks that these businesses are "bad" and that if they provide service to them, they'll be audited from top to bottom. It is a defacto [washingtonpost.com] suppression of Free Enterprise b
Re: (Score:2)
Forcing companies to provide mortgages to people who are patently unqualified is an example of unreasonable regulations that resulted in untold devastation to the economy.
Nobody EVER did that. They were required to be a bit more flexible in determining qualifications for loans on starter homes (that is, small mortgages) and to stop blatantly racist redlining. Instead, they began actively talking people into huge loans with built in time bombs for McMansions (that is, huge mortgages). Then they invented a variety of wildly complex new financial instruments based on those crazy loans deliberately (and fraudulently) mis-represented as AAA investments.
In other words, the banks a
Re: (Score:2)
Banks were implicitly threaten with audits if their loan profiles didn't meet certain expectations. It's the same with Operation Choke Point today where they are told they will probably be audited if they don't conform their client list to excluded businesses that match certain profiles.
Don't confuse explicit demands with implicit threats, which can be just as effective in controlling behavior.
Re: (Score:2)
The conditions on the implicit threats would have easily been met by complying with the law and making more loans on STARTER HOMES to people who they might otherwise redline.
Don't fall for the flimsy excuses.
Re: (Score:2)
Just like stores banned negros because their 'careful informal studies' (read prejudice) told them negros were more likely to steal.
I notice that when stores were forced to de-segregate they had enough sense to not invite their new customers to walk out with the cash box 'just to be sure they weren't fined'
Of course, since they firmly believed those loans were likely to end in a default, why did they bundle them up and sell them as AAA financial instruments?
Re: (Score:2)
On the other hand, I've found that people talking about explicit demands are more reliable than ones talking about implicit threats. People tend to find concepts that support their ideology, and it's always easy to find implicit threats when you want to.
Re: (Score:2)
Nice mortgage business you have there. It'd be a shame if you were to be labeled as some racist organization or something.
Re: (Score:2)
Now, if you can back that up, so it doesn't look like you're typing that under an implicit threat, that'd be good.
Re: (Score:2)
And, of course, the bigger the loan, the more they could sell it for. Turning down loans wasn't profitable at all.
Re: (Score:2)
What the hell are people talking about quiet before the storm ?
What do you mean foundations have been laid ?
Bunch of BS. Large companies are starting over, without the legacy:
https://www.youtube.com/watch?... [youtube.com]
Re: (Score:2)
Foundations - in the way that now there's optical fiber backbones all over the place, and there's a proof that it is feasible to offer internet services to most citizens.
Re: (Score:2)
I guess it depends on the type of foundation you mean.
Things like fiber have been there for many, many, many years now. I hardly see it as anything new. Or anything part of this quiet, they talk about it.
There are some big changes coming in fiber optics though. Silicon photinics.
Re: (Score:2)
Fiber would be something new in my neighborhood.
Re:Horseshit (Score:5, Insightful)
alternately, it will soon be time for the pendulum to swing back to "we've got to have everything in-house, these security breaches are killing us" and "dumb terminals and having everything in the 'cloud' is killing productivity when the cloud is down, we need real apps so users can work even when the cloud doesn't"
Re: (Score:2)
Re: (Score:2)
If the "transcontinental railroad" is truly built, then the cloud won't be going down (for any significant amounts of time) in the future.
How often do you venture out onto the Eisenhower Interstate Highway system, stymied that you can't use it in the normal fashion (yes, rush hour in metro areas still needs work, mostly population control, I say, but...)?
If your "cloud is down" more than 5 minutes per day, or has a big (multi hour) outage more than once a year, then you have not yet arrived at modern (2014)
Re: (Score:2)
Re: (Score:2)
The railroad is built, and kicked butt over ox and wagon, and has been outmoded by car/truck (thus my example) and plane - and left to rot.
Yes, train service sucks - unless you are a hopper car full of ore and don't particularly have a schedule.
Re: (Score:2)
Centralize/Decentralize YAWN. I've seen 3 swings in 20 years of pro, myself.
A rather simplistic hardware-centric view (Score:5, Informative)
The article is a rather simplistic hardware-centric viewpoint. It doesn't even begin to touch on the areas where IT has always struggled: design, coding, debugging, and deployment. Instead it completely ignores the issue of software development, and instead bleats about how we can "roll back" servers with the click of a button in a virtual environment.
Which, of course, conveniently ignores the fact that someone has to write the code that runs in those virtual servers, debug it, test it, integrate it, package it, and ship it. Should it be an upgrade to an existing service/server, add in the overhead of designing, coding, and testing the database migration scripts for it, and coordinating the deployments of application virtual servers with the database servers.
Are things easier than they used to be? Perhaps for they basic system administration tasks.
But those have never been where the bulk of time and budget go.
Re:A rather simplistic hardware-centric view (Score:4, Interesting)
Indeed, and virtualization is a rapidly evolving part of infrastructure right now. We may no longer be upgrading the hardware as rapidly (although I'm not certain about that either), but the virtual layer and tools are changing, and upgrading those requires just as much upheaval.
Re: (Score:1)
Want a new dev environment? *click*
Want a new db server? *click*
Need an FTP server? *click*
Need an HTTP server? *click*
Before you know it when you need to deploy a small software change it becomes a big deal because you have a billion bloody servers to update.
Before virtualization (or at least the ease of virtualization) you took your time and planned - checked available resources etc. Resources were scarce, RAM wasn't so abundant,
Re: (Score:2)
Really?! What about all the data and the configuration?
I suppose your *click*, *click* includes all the steps in a parallel deployment.
Hardware still matters (Score:2)
It seems too many forget that all this virtualization still runs on physical servers. Those physical servers still need hardware upgrades, monitoring, and resource management (especially when one starts oversubscribing). I don't get why people keep thinking hardware went away. Instead of lots of 1U servers, now you have big iron running lots of virtual servers.
Re: (Score:2)
Yes, but now you have one, maybe two (hopefully super-smart) guys onsite with a deep systems knowledge, instead a fleet of screwdriver wielding guys with an A+ certification who are as likely as not to screw up your system. Once it's up and running you just have to keep that machine and it's backup going, and everyone can build on top of that in software, from anywhere in the world.
Re: (Score:2)
Hardware does still matter, but it's no longer something that must be watched closely and in fear.
Re: (Score:2)
Of course it didn't go away, but it did get a lot easier to maintain.
I remember very well back in the bad old days, that white knuckle time between telling the remote server to reboot with a new kernel and ping starting again. And of course, the advance setup where you make the old kernel the default in hopes that if it all goes sideways you can call and find someone on-site who can manage to find and press reset should it hang or have some random problem that keeps it off the net.
Then came nicer setups whe
If you didn't know what you were doing ... (Score:5, Informative)
They could be if you did not know what you were doing. Like I suspect the author of TFA did not know.
From TFA:
If he's talking about a production system then he's an idiot.
If he's talking about a test system then what does it matter? The time spent running the tests was a lot longer than the time spent restoring a system if any of those tests failed.
And finally:
WTF is 10Base-2 doing there? I haven't seen that since the mid-90's. Meanwhile, every PC that I've seen in the last 10 years has had built-in gigabit Ethernet.
If he wants to talk about hardware then he needs to talk about thing like Cisco Nexus. And even that is not "new".
And, as you pointed out, the PROGRAMMING aspects always lag way behind the physical aspects. And writing good code is as difficult today as it has ever been.
Re: (Score:3)
WTF is 10Base-2 doing there? I haven't seen that since the mid-90's.
That was probably the "or so" that came after the word 'decade'.
A "decade or so" could be taken to mean 2 of them, which puts it back in the mid 1990's.
I think you're just being Captain Pedantic, when all the GP was really trying to say was that things move pretty quickly in IT.
Re:If you didn't know what you were doing ... (Score:5, Insightful)
Re: (Score:3, Insightful)
Re: (Score:2)
> There is another trend of third parties marketing infrastructure solutions to high level management, skipping local subject matter experts.
Trend? I thought this was SOP for sales, because SMEs kill sales. This has been known by Salesmen and SMEs since time immemorial.
Re: (Score:2)
If he's talking about a production system then he's an idiot.
Why? Is it your contention that the work of sysadmins and support personnel has just been trouble-free for decades, and all the problems were caused by a sysadmin "not knowing what they were doing"?
Re: (Score:2)
WTF is 10Base-2 doing there? I haven't seen that since the mid-90's. Meanwhile, every PC that I've seen in the last 10 years has had built-in gigabit Ethernet.
Right, within a decade, the '90s to be specific, all of those network technologies were in use at some point.
Re: (Score:2)
Meanwhile I'd like to be able to turn clusters into a virtual server instead of having to codes specificly for clusters. Something like OpenMosix was starting to do before it imploded. Make serveral machines look like one big machine to applications designed to only run on single machines.
Re: (Score:3)
www.scalemp.com does what you request.
It's not exactly all warm and fuzzy. Things are much improved from the Mosix days in terms of having the right available data and kernel scheduling behaviors (largely thanks to the rise of NUMA architecture as the usual system design). However there is a simple reality that the server to server interconnect is still massively higher latency and lower bandwidth than QPI or HyperTransport. So if a 'single system' application is executed designed around assumptions of n
Re: (Score:2)
Re:A rather simplistic hardware-centric view (Score:5, Insightful)
Even software is slowing down, though. A lot of the commodity software reached the point of 'good enough' years ago - look how long it's taken to get away from XP, and still many organisations continue to use it. The same is true of office suites: For most people, they don't use any feature not present in Office 95. Updating software has gone from an essential part of the life cycle to something that only needs to be done every five years, sometimes longer.
Around 2025 we will probably see a repeat of the XP situation as Microsoft tries desperately to get rid of the vast installed base of Windows 7, and organisations point out that what they have been using for the last decade works fine so they have no reason to upgrade.
Re: (Score:2)
Updating software has gone from an essential part of the life cycle to something that only needs to be done every five years, sometimes longer.
Not if you actually care about security. The older software and operating systems prior to Windows 8 don't support newer more-effective security attack mitigation approaches such as ASR.
Re: A rather simplistic hardware-centric view (Score:1)
By then there'll be virus scanners in switches
Re:A rather simplistic hardware-centric view (Score:4, Insightful)
I'm talking about new product major versions, not just patches.
The only reason many organisations are ditching XP right now is that MS stopped supplying updates. That isn't "Getting new software to further advance the organisation." That's more "Reluctantly going through the testing and training nightmare of a major deployment because Microsoft want to obsolete our otherwise-satisfactory existing software."
Re: (Score:3)
A lot of the commodity software reached the point of 'good enough' years ago - look how long it's taken to get away from XP, and still many organisations continue to use it.
I find it hard to believe that operating systems became "good enough" with Windows XP. Rather, Vista took so long to come out that it disrupted the established upgrade cycle. If the previous 2-to-3-year cycle had continued, Vista would have come out in 2003 (without as many changes, obviously), Windows 7 in 2005 and Windows 8 in 2007. We'd be on something like Windows 12 by now.
It's good that consumers are more aware and critical of forced obsolecence, but I don't agree with the "XP is good enough" crowd. I
Re: (Score:1)
Well beyond hardware, Software reliability over the past few decades has shot right up.
Even Windows is very stable and secure. Over the past decade, I have actually seen more kernel panics from Linux than a BSOD. We can keep servers running for months or years without a reboot. Out Desktops,Laptops, and even mobile devices now perform without crashing all the time, and we work without feeling the need to save to the hard drive then backup to a floppy/removable media every time.
What changes has happened se
Re: (Score:2)
Software reliability over the past few decades has shot right up.
I think this is a questionable premise.
1) Accurate, though has been accurate for over a decade now
2) Things have improved security wise, but reliability I think could be another matter. When things go off the rails, it's now less likely to let an adversary take advantage of that circumstance.
3) Try/Catch is a potent tool (depending on the implementation it can come at a cost), but the same things that caused 'segmentation faults' with a serviceable stack trace in core file cause uncaught exceptions with a
Re: (Score:2)
The article is a rather simplistic hardware-centric viewpoint. It doesn't even begin to touch on the areas where IT has always struggled: design, coding, debugging, and deployment. Instead it completely ignores the issue of software development, and instead bleats about how we can "roll back" servers with the click of a button in a virtual environment.
And now is when we have a long and stupid debate as to whether the term "IT" signifies a grouping of all computer-related work including development, or whether it's limited to workstation/server/network design, deployment, and support. And we go on with this debate for a long time, becoming increasingly irate, arguing about whether developers or sysadmins do more of the 'real' work, and...
Let's just skip to the end and agree that, regardless of whether IT 'really' includes software development, it's pret
Re: (Score:3)
Re: (Score:2)
The article starts with the observation that the hardware bottleneck is mostly gone, if you can afford to supply basic coffee to your employees, the IT hardware doesn't cost much more than that - contrast that to 1991 when the PC on my desk cost 2 months of my salary, and our "network" was a 4 line phone sitting next to it (modems came to our office 5 years later).
Then, let's dream about what's next... you can dream, can't you?
Re: (Score:1)
Not paying attention? (Score:5, Interesting)
8k video? (Score:1)
Re: (Score:1)
By then we will have the infamous Sloot Digital Coding System that will encode an entire movie down to 8KB so what are you so worried about?
http://en.wikipedia.org/wiki/Jan_Sloot
Re: (Score:2)
The only problem there is that it is, for most purposes, pointless. Most people would be hard pressed to tell 720p from 1080p on their large TV under normal viewing conditions. What we see really is a placebo effect, similar to the one that plagues audiophile judgement: When you've paid a heap of cash for something, it's going to sound subjectively better.
Re: (Score:1)
Re: (Score:1)
It's easy to tell the guys from IT. Everybody else reads the PDF files for their content. The IT guy looks for artifacts in the font rendering.
Re: (Score:2)
Exactly, yet you will have a TON of people claiming they can tell the difference. In reality they can not.
99% of the people out there sit too far away from their TV to see 1080p Unless they have a 60" or larger and sit within 8 feet of the TV set. The people that have the TV above the fireplace and sit 12 feet away might as well have a standard def TV set.
But the same people that CLAIM they can see 1080p from their TV 10 feet away also claim that their $120 HDMI cables give them a clearer picture.
Good lord (Score:1)
Absolutely (Score:1)
Good (Score:2)
Now standardize all your password requirements to a strength-based system without arbitrary restrictions or requirements, and standardize your forms' metadata so that they can be auto-completed or intelligently-suggested based on information entered previously on a different website. Trust me, this sort of refinement will be greatly appreciated.
What about security (Score:1)
I don't subscribe to this rose-tinted point of view, especially if you look at all this beautiful tech from the security standpoint.
Most of the tech we deal with today was originally designed without security concerns. In most cases, security is an afterthought.
So much for sitting back and taking a break.
Re: (Score:2)
Yeah, "... In essence, we have finally built the transcontinental railroad, and now we can use it to completely transform our Wild West."
There are tunnels filled with dynamite all underneath the track just waiting for some wild west yahoo to push the detonator.
Once a month, Microsoft issues security blankets in an effort to hide them.
Until that problem is solved at the roadbed, no one's going to get a good night's sleep because those bird whistles you hear are not authentic, Kemo Sabe.
Ah yes (Score:2)
Moore's law has run out of steam. Yay!
Re: (Score:2)
I think it's more about the end of the MHz wars. Nowadays, to get more power, you add more cores. If you can't do that, you add more boxes.
If you've got a single threaded million instruction blob of code, it's not executing very much faster today than it was a few years ago. If you're able to break it into a dozen pieces, then you can execute it faster and cheaper now than you could a few years ago, though.
Moore's law hasn't really run out of steam, it more that it's rules have changed a bit - the raw power
Re: (Score:2)
you can have 256 cores at 500mhz and a 4 core 5ghz system will be a lot snappier and faster. because most of what is used for computing does not scale to multi core easily.
I will take a 2 core 5ghz system over a 4 core 3ghz system any day.
Re: (Score:2)
Moore's law was about a one node shrink every 18 months, meaning a reduction in structure sizes by sqrt(2), i.e. twice the number of transistors at the same die size. The reduction in size meant a reduction in gate thickness and operating voltage by sqrt(2) and a reduction of capacitances by a factor of two. Those allowed an increase in clock speed of sqrt(2) at constant power. None of that is happening any more.
Wrong (Score:2, Insightful)
Re: (Score:3)
Re: (Score:1)
With the increased reliability of modern cars, people do make fewer trips to the garage. So it's not unlikely that cars won't be in the garage more than every 5 years.
I guess the same fact being true for IT really bugs you. The IT drones where I work are right now in a tizzy because the corporate IT people in Mexico are taking over. Because they can, and it saves a lot of $$, and also because the local fucks just aren't needed much anymore. There's no need for a guy to clean the lint out, all the mice are o
Re: (Score:2)
Re: (Score:2)
Re: (Score:3)
Oh, wait, you've already mentioned you're a business person. Enjoy your Dunning-Kruger while it lasts.
need to fix the infrastructure when it are broken.
Shall we fix your understanding of the English language while we're at it? Or would that be too mission-critical a business decision?
Re: (Score:2)
Re: (Score:1)
Sure thing, massa! Nice private Emails and interesting browsing behaviour over there.
How much to keep my mouth shut? Looks like I might earn more $$ than you, after all, dear BA boy.
Oh BTW, the toner is replaced -- oh, look, it's already printing. Wait, why do you...?
Oh, neat. Looks like you're about to be fired.
Re: (Score:3)
Awesome, we need to join the plumbers union and start getting $125 an hour then. Thanks for your support!
Oh lookie... (Score:4, Informative)
Re: (Score:2)
These are the sorts of things stupid people say... (Score:1)
...right before the next, undreamed-of computing revolution knocks everyone on their ass.
Re: (Score:1)
Looking at the past... (Score:5, Insightful)
You, sir, are a buggy-whip manufacturer (as well as a dinosaur).
Re: (Score:3)
Seems like you are talking about general computing and related applications of computing. This guy is talking about business IT, which is a tiny subset of computing.
It would not be inappropriate to mod parent off topic for posting a generalist reply to something written for a specific audience.
Re: (Score:3)
"the technologies...have stabilized" (Score:1)
What the -- ?? "the technologies behind our servers and networks have stabilized" -- when did this happen? I'm not a datacenter person, but isn't the world filled with competing cloud providers with different APIs, and things like OpenStack? Did all this stuff settle down while I wasn't paying attention?
Another layer has solidified (Score:2)
I think would be a better way of looking at what this article is on about.
Back in the late 80's early 90's when I graduated and started my career in the Networking Industry the OSI 7 layer model (https://en.wikipedia.org/wiki/OSI_model) was often referred to. You don't hear it mentioned much these days.
If you applied IT history and economics to it you'll find that each of those layers saw a period of fantastic growth & innovation (a few short years) before becoming IT commodities and having little valu
um (Score:3)
Now that the technologies behind our servers and networks have stabilized, IT can look forward to a different kind of constant change, writes Paul Venezia.
I don't think Paul Venezia works in IT.
Comment removed (Score:3)
Like the railroad, its about control (Score:1)
Translation: Bandwidth and ubiquitous connectivity, along with a generation trained to have no privacy are in place. Let the police state begin.
If you think things like rural electrification are about helping people, you have your head in the sand.
IT Revolution? (Score:1)
At the risk of pissing off some folks, I must say I've worked in IT since before it was called IT, and I can honestly say no revolutions will come from that area. After all, IT isn't known for it's innovative and R&D atmosphere. IT is the result of cramming middle management, contractors, and novice-to-mediocre developers together in cubicles. Sure, it's a steady paying job, which is why most of us do it. The revolutionary stuff will continue to come from those who have the luxury of choosing not to
Re: (Score:2)
Quiet != stasis (Score:2)
The on-going technology churn we've seen in the last decade is *not* a feature of a revolution in progress that may be coming to an end, it's a reflection of stagnation in technology, without the ideal data centre technology (at least in terms of software) having achieved any kind of dominance. There's been a endless parade of new web technologies, none of which is more than an ugly hack on HTML. Websites are better than they were in twenty years ago, but certainly not 20 years' worth of progress better.
Nothing has stopped or stablized (Score:3, Interesting)
The concept is false. Things have changed in how they break and what we are concerned about on a daily basis. 10 years ago I didn't have compromised accounts to worry about every day. But I did spend more time dealing with hard drive failure and recovery. We are still busy with new problems and can't just walk off and let the systems handle it.
If you believe IT is like running your Android device, then yes, there is little to be done other than pick your apps and click away. If you have some security awareness you would know there is much going on to be concerned about. When the maker of a leading anti-virus product declares AV detection is dead, it is time to be proactive looking at the problem. Too many IT folk believe if there is malware it will announce itself. Good luck with that assumption.
Re: (Score:2)