Competition is decreasing in enterprise IT – and you’ll be poorer and dumber for it HPE’s decision to acquire Juniper is bad news for enterprise IT, as yet another example of consolidation in a field that already offers fewer, and less palatable, competitive choices in a shrinking market. Networking is already a difficult area: Cisco’s $57 billion annual revenue vastly exceeded that earned by its main …
These big companies can do this, because they tend to over charge customers and under pay their workers.
Sure, you may say workers earn relatively good money in comparison to other workers, in other professions, but remuneration is not proportional to the value they create for the company.
The companies sit on piles of cash that otherwise should have been paid to workers as a salary.
Just like the prices charged by companies to customers for their products, the salary earned by workers is not determined by the value they bring, but by how hard they are to replace.
The price you pay for a server is not proportional to the money you'll make with it. The salary of an employee is not proportional to how profitable they are to the company. Miners earn about the same in a coal mine or a diamond mine.
...but the investment will always flow to the diamond mine because the investors just look at the potential for $$$$$$ returns. Not the fact that, if things go South, you can't burn diamonds to keep the lights on and your feet warm.
"But with on-prem enterprise IT now a shrinking market, investors and innovators alike will be wary."
It's only a shrinking market until customers realise they might be better off on-prem. What's described here is not dissimilar to the days of mainframe dominance. Those days were ripe for disruption which happened when PCs were invented. Now we don't even need the invention step, the PCs are already here.
> It's only a shrinking market until customers realise they might be better off on-prem.
The enormous complexity and high cost caused by running and maintaining complex setups of traditional enterprise software on-prem were what led most customers to start investigating cloud in the first place. Cloud was cheap for a while, so instead of getting complexity and volume under control on-prem, the exiting hairball of code was thrown into the cloud.
Today, there is no longer a cheap and secure platform to run a lot of complex stuff, so customers IMHO need to do it it the hard way this time:
Start getting it (more) simple again, reduce the number of activce software components. Reduce complexity. Reduce use of traditional "Enterprise" catch-all software that often lump dozens of separate tools together to create more complexity than they solve.
From there deploy a lower number of potentially simpler components either on-prem or in cloud (probably a mix will be optimal) of what you really need.
The enormous complexity and high cost caused by running and maintaining complex setups of traditional enterprise software on-prem were what led most customers to start investigating cloud in the first place.
This has been mostly created by fear.
fearmonger> What if you suddenly get a million visitors and your page stops responding?
This was true 20 years ago. You had to have some set up to accommodate spikes of traffic, because machines back then were not up to it.
Cloud was a promise that you can add resources with a click of a mouse in such an event and that was taking the anxiety away.
But for most companies such event never happened.
Today a single server can most likely handle all traffic a typical business would have (like a few million customers level). There is no need for complex setups.
You may add extra machines for content and redundancy, but these things are not complex anymore.
fearmonger> What if your server crashes? In the cloud this never happens and if it does, you can recover with a click of a mouse
It's probably more likely you misconfigure your cloud or a newbie deletes important instance than your on-prem server dies.
Cloud can go down too, but then if it does you are not really in control when it happens. You have to wait and hope you are important enough for someone to look into it.
fearmonger> You need to have someone babysit your infrastructure.
Typically true, but cloud is so expensive, the savings from on-prem could let you hire people looking after it 24/7 easily and difference is they'll care about it.
fearmonger> Everyone uses cloud. If you don't then you are doing it wrong.
Everyone uses drugs...
So all I need to do is write some software that runs on a white label box, then instead of paying a hyperscaler rent for my servers, I get the customers to hold them on site, and push updates. No admin for customer as it’s done remotely, just as I would in the cloud, but they get a discount and alll data locally stored if that’s their bag
The thing is, the VC funds and other investors have their money invested in these cloud companies and typically the requirement for any company they want to invest in is to have their services run in the cloud.
If they run on prem, they will be told they won't get the investment. It's a huge conflict of interest I don't believe has been investigated or tackled in any way.
Even if you demonstrate the savings and show the benefits over the cloud, you'll get hard no from most of the start ups.
There is a reason they do that when they do, which I think you've overstated. They're requesting it because VCs tend to want the companies they invest in to scale really rapidly. Sometimes, that's a viable approach, sometimes not, but the VCs generally expect it. This means that, if you get a hundred times as many users next month as you have this month because you spent the money on a very successful awareness campaign, the VCs really want you to be able to handle that. If your infrastructure isn't up to the demand, it takes a long time to scale it up. You have to get equipment, configure it, and set it up, but perhaps more importantly, you need the space to store it, the ability to power and cool it, and those things take a long time to expand for a startup that probably has one small office, not their own private datacenter. Even colos won't be able to instantly provide a ton of space overnight.
A startup that can convince VCs that they can scale up without having to buy a bunch more infrastructure can probably get around this, but if that's not something the startup techs are willing to promise, then the investors will ask about using cloud for scaling potential.
It's rare that a VC would invest in an established business - not unheard of, but invariably quite rare. They want to invest in startups, get in on the ground where the opportunity to maximise return on their investment is at its highest. At this point they're interested in costs - and making a big capital investment on infrastructure would undoubtedly be frowned upon, they'd much rather see initial costs start low and ramp up on a PAYG basis.
It's only a shrinking market until customers realise they might be better off on-prem
And they look at it and find ... that some of the key vendors no longer offer on-prem as an option. We're rapidly heading towards a world where you can't buy (or rent*) an on-prem licence so your only option will be the vendor's cloud. MS have the biggest stakes here since (by well documented illegal activities) they engineered themselves a de-facto monopoly on business desktops, and they've made darn sure that people go with the rented cloud rather than perpetual on-prem. I recall them trying this back in the 90s but the world wasn't ready - and businesses were rightly wary of getting into the "keep paying or you lose access to your business". I'm not entirely sure what's happened to change that.
* In a world where perpetual licences are no longer available, you rent your software and get evicted if you stop paying the rent. Just like you get evicted and lose your business if you stop paying the rnet on the factory.
Operating model:
1) Brownnose the target's execs, and over free lunches and "fact finding" trips, tell them how skilled BigCo is, how much unbelievable talent it has at its fingertips, how much value it creates for its customers.
2) Tout 20% (that's the magic number) savings compared to the target's cost of tech/services. "That 20% mounts up year on year, but don't forget how much value we can create through our process expertise!". Knowing what 20% is doesn't need inside knowledge, the cost of every operation is well understood in the world of services and IT management. In reality there are no savings when BigCo do it - any target large enough to be of interest will have have its own economies of scale, a skilled workforce, competent IT, and functional management who understand full well that they're an overhead to be kept as low as possible. But target directors aren't bright enough to know this.
3) When the target invites bids, bid in at 25% below the anticipated in-house cost. Nobody will query the bid numbers, or the operating model. All the time, BigCo's capable and well rewarded sales and legal teams will be running rings around the target's beleaguered, outnumbered, out-experienced, and under-rewarded in house teams.
4) Target TUPE's across it's own team and processes, in return for a 5 year deal. For the first 14 months, other than teething troubles it all seems good. Up to this point, Big Co has been making a loss, because it's had the in house team, and been paying (the least possible) to whittle them down, and even when it's 100% offshore wage slaves, the inherent inefficiencies of offshore plus pretty high account management costs and bid costs mean their routine operating COST is near enough the same as the old in house team.
5) Now BigCo starts to progressively ramp the invoices - not just for the losses made to date, but because unlike the in house team, BigCo need to make a big fat 30% gross margin. This is normally achieved through things like non-standard service charges, and the inevitable changes that any corporate makes in a few years. So new divisions to serve (or not serve if disposed of), new processes, new laws or regulatory processes, new accounting standards - everything that the customer forgot to include, or couldn't specify, that's where the moolah is.
6) Four years in, the customer is paying 30% more than they were when they did it themselves. But by then all the senior management have swapped seats, the original business case and its assumptions is long, long forgotten, and there's no in-house team to say "Hold on! We're being ripped off!"
7) Target company directors pay themselves a big bonus, slap each other on the back in congratulation at what mega-business brains they are, because they've "simplified their processes", "engaged a world leading business partner", "got access to the unique skills/AI prowess of BigCo".
8) Target company investors wonder why costs have gone up. Target company employees wonder why the service has turned to shit.
Meanwhile the merry-go-round of C-level dirtbags goes on and with each revolution they get fatter and fatter bank accounts (offshore naturally).
Meanwhile the workers get fed peanuts for promises and are left to eat gruel.
Eventually, the magic of cloud and headcount reduction/rightsizing/downsizing will get to such a level that the business will suffer hemorrhaging sales and all those stock options will mean zip.
The business will go bust and be sold to the Chinese.
As Theodore Sturgeon said, “99% of everything is crud”.
Software-defined networking and white-box hardware doesn't seem to have disrupted the market, as hoped by some.
They could have achieved quite high levels of adoption already, but those installations are essentially invisible because of their low revenue impact. Smart folk are using them, and not making a big fuss about it. Because there aren’t big news headlines about it, the trend-followers don’t notice. They stick to the high-cost, low-quality conventional products from Cisco and the like.
And that suits the smart folk just fine.
> To those who want to sell proprietary products—no point at all. It’s something they would wish out of existence if they could. To prospective users, who would rather not pay lots of money to those vendors of proprietary products—all the point in the world.
What do these SDNs run on - virtual software vapor ?
All the Marxist state leaders in the world envy the CCP. This is the reason for industry consolidation. Following the CCP model of controlling an industry is easier than letting "free market capitalism" dictate the markets. There is no communist control in what you "can't" control.
This is not limited to the IT industry as other industries have also consolidated. How many phone companies are left(well I guess these are IT companies)? How many major airlines are left, how many oil companies are left, and how many, etc... etc... Look 20-30 years ago. There were a lot more companies in these industry sectors back then.
Of course, this is just my observations/opinions and I hope I'm wrong.
"Big Tech wants you to think that’s in your best interests, because you no longer have to develop intimate understanding of tech."
That isn't always the case - configuring some of these cloudy services (thinking IaaS especially) is no more straightforward than configuring on-prem, and sometimes it's actually more complicated.
The problem we have now is that Big Tech is really trying to force the issue by putting obstacles in the way of running things on-prem.
Surpise surprise - the sharp practices of companies such as CISCO to initially allow license transfer/extension/discount for SDN appliances in the cloud have ensured they retain their customer and future-proof their cash-cow (very similar to what Microsoft are being investigated for with their License Portability for Windows Server and SQL Server).
I guess it takes a brave organisation to 'throw the baby out with the bathwater' and break their reliance on the likes of CISCO, and for every one of those 'brave' organisations, there are thousands of others not quite so brave!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
