Cringley's Next 2019 Predictions: Only 3.5 Cloud Players Will Survive (cringely.com) 148
Ten days ago 66-year-old tech pundit Robert Cringely revealed the first of what may be his final set of annual predictions for the technology industry -- but he's not done yet. Thursday Cringely predicted that "the Virtual Private Cloud (VPC) solution based on Open Source using Linux will change the Internet-as-a-Service Cloudscape to VPC-only during 2019" -- and that there'll be an industry-wide shakeout.
Long-time Slashdot reader supremebob, a Connecticut-based sys-admin, writes: He seems to believe that IBM Cloud and Oracle Cloud and doomed to fail, and Alibaba will only survive because of its strong Chinese presence. These seem like safe predictions, but his comments on Google Cloud are somewhat controversial...
After AWS, Alibaba, and Microsoft, "All the others will eventually disappear," Cringely writes, adding "Remember you read it first here." Google's largest cloud customer will always be Google and that will inevitably lead to poorer service for outside customers. That's why I think of Google Cloud as half of a player. Feel free to prove me wrong by delighting customers, Google... I don't see the marketing effort to help clients migrate. Lots of handholding is needed that IBM and Microsoft are happy to provide. Google does not understand customers whose IQs are sub-200. As such, Google doesn't have (and likely won't) have a history of winning outside of search advertising.
For IBM, their VPC roll-out is coming in the next month or two, but it's more marketing than an actual product. Big Blue simply has no capital to build out a unique offering. And Oracle? Well the new head of Google Cloud came from Oracle, where not enough was happening.
Cringely also predicts the U.S. government will try to force Amazon to spin-off its near-monopoly cloud business, noting that "the larger customers of AWS (those not operating on a credit card) generally hate Amazon because of its ruthless business behavior."
Lots of pressure will come to bear in this case from IBM, Microsoft, and Oracle, who are all suffering from a very specific database problem competing with AWS. Each of these companies sells their own database (DB2, SQL Server, and Oracle, respectively) that they've rolled into their cloud services. AWS's RDB, in contrast, is based on MySQL and costs Amazon almost nothing to support, giving the biggest cloud player a clear pricing advantage.
Long-time Slashdot reader supremebob, a Connecticut-based sys-admin, writes: He seems to believe that IBM Cloud and Oracle Cloud and doomed to fail, and Alibaba will only survive because of its strong Chinese presence. These seem like safe predictions, but his comments on Google Cloud are somewhat controversial...
After AWS, Alibaba, and Microsoft, "All the others will eventually disappear," Cringely writes, adding "Remember you read it first here." Google's largest cloud customer will always be Google and that will inevitably lead to poorer service for outside customers. That's why I think of Google Cloud as half of a player. Feel free to prove me wrong by delighting customers, Google... I don't see the marketing effort to help clients migrate. Lots of handholding is needed that IBM and Microsoft are happy to provide. Google does not understand customers whose IQs are sub-200. As such, Google doesn't have (and likely won't) have a history of winning outside of search advertising.
For IBM, their VPC roll-out is coming in the next month or two, but it's more marketing than an actual product. Big Blue simply has no capital to build out a unique offering. And Oracle? Well the new head of Google Cloud came from Oracle, where not enough was happening.
Cringely also predicts the U.S. government will try to force Amazon to spin-off its near-monopoly cloud business, noting that "the larger customers of AWS (those not operating on a credit card) generally hate Amazon because of its ruthless business behavior."
Lots of pressure will come to bear in this case from IBM, Microsoft, and Oracle, who are all suffering from a very specific database problem competing with AWS. Each of these companies sells their own database (DB2, SQL Server, and Oracle, respectively) that they've rolled into their cloud services. AWS's RDB, in contrast, is based on MySQL and costs Amazon almost nothing to support, giving the biggest cloud player a clear pricing advantage.
Re: Pretty sure google will be around forever. (Score:2, Insightful)
Google Public Cloud will be around forever as a 0.5 Cloud and you will lose your competitive edge because Google will prioritize their needs over yours. All this, according to the summary.
Re: Pretty sure google will be around forever. (Score:2)
Re: (Score:2)
Google Cloud seems to have the most polished services for hosting containers at this point, which makes sense considering that they helped to develop the technology.
They also have partnerships with other major hosting providers like Rackspace and Salesforce, which makes me also think that they will be around for awhile.
Re: (Score:3, Insightful)
You'd think.. except, they may get bored, and just casually mention that they're going to shut down in 3 months... please transition elsewhere.
This is the stability one should expect from Google, based upon real, historical data.
Re: (Score:2)
"I figure a couple bigger players in cloud systems will get hacked or somehow compromised or simply have a major failure losing lot's of client data and that will be the reason many will back away from the cloud."
Ha!, Ha!, and then onether Ha!
Of course a couple big players will get hacked, of course they will lose lots of client data but, no, that won't make business going away from the cloud. There will be, if any, a minor "glitch" on one cloud providers of customers moving to a different one.
On one hand,
Re: Big players will get hacked (Score:3)
Re: (Score:2)
Don't question what Oracle provides, simply pray they do not alter the deal further.
IBM isn't putting out a "unique offering" because they're focusing on using generic standards-based tools, and selling the full corporate servicing you mention.
As a consultant, I want to use the same technologies that IBM is using, and I want to be able to tell my clients, "If you outgrow my services, you can go to IBM and keep using the same stuff you started with."
That's way better marketing than, "If you outgrow my servic
Re: Big players will get hacked (Score:2)
Internet-as-a-Service (Score:2)
WTF is Internet-as-a-Service? As-a-Service usually means it's somewhere else and you have to give somebody a load of cash and use the internet to reach it.
Do I need to buy more internets now?
Re: Internet-as-a-Service (Score:1)
Re: (Score:2)
With internet as a service, you would never even worry about buying more internets; you'd have all the internet you needed right at your fingertips!
I guess the alternative to internet-as-a-service would be to lay some fiber to the backbone and try to sign a peering agreement with somebody?
Re: (Score:3)
In once sense, the "internet" is simply a lot of interconnected networks. Following? AWESOME! Now, there is a new "as a service" that is emerging right now, SD-WAN, or "software defined wide area network" - it is simply a virtualized network on top of the internet, much like VLANing is a virtualized network on top of a physical network. SD-WAN is essentially an "internet as a service" as it allows multiple networks to become interconnected virtually to create a wide-area, or pseudo-internet on top of the ph
Re: (Score:1)
Re: (Score:2)
SD-WAN "competes" with MPLS in the sense that the vendors want you to ditch your expensive MPLS links for their offering that runs on commodity internet bandwidth. They do offer some decent features (everything after the first couple of hops is typically private network, so latency is fairly predictable/manageable/comparable to MPLS), they can bake in WAN acceleration, redundancy is trivial, etc.
In the end, though, they cost as much (or more than!) MPLS. Typical cloud business model of replacing what you h
Re: (Score:2)
OMG, someone invented a router with a VPN connection. Or maybe they just use SSH tunnels?
Yo dawg - i herd u like internets (Score:2)
XaaS - Xzibit as a service?
Re: Cringley is a moron (Score:5, Insightful)
AWS has the right tool at the right price. Why go anywhere else?
You might think differently about that if Amazon starts competing in your industry.
I do agree that at least AWS should be split-off from the rest of the business.
It's bad enough that we have banks that are "too big to fail". We don't need an IT service provider in the same league.
Re: (Score:1)
The pending orders against Microsoft, and the threat of breakup turned out to change Microsoft's behaviour enough to allow for the market to succeed. And it was 20 years ago, not 30.
Re: (Score:2)
AWS is limiting and expensive for people who can set up their own systems. AWS gives you push button access to a subset of what's possible and that button is expensive enough to rule it out as an option for businesses doing organic growth.
Re: Cringley is a moron (Score:5, Interesting)
Re:Cringley is a moron (Score:5, Interesting)
Funnily enough "might as well give all your IP to the Americans" is what the Chinese think about AWS, and also that being at the whim of the US government could be bad for your reliability. Being Chinese, they're already at the whim of the Chinese government - why add trouble by using a US cloud provider?
Re: Cringley is a moron (Score:2)
Re: (Score:2)
Also documentation and support available in Chinese. Alibaba is very cheap as well.
Re: (Score:2)
Also documentation and support available in Chinese. Alibaba is very cheap as well.
Is it connected to the Internet too? Or only to the Chinanet (you know, that country-wide intranet thing that also features a couple of heavily firewalled gateways to the Internet)?
Re:Cringley is a moron (Score:5, Interesting)
AWS, yes. IBM and Oracle, no. Sure, they may always exist, but they will likely never be large enough to be relevant. Neither has shown any sort of innovation in that space and are just me-too'ing it. And for Oracle it's doubly bad because they have such a terrible reputation. MSFT is far bigger and entrenched than IBM or Oracle.
In comparing the various sizes of these providers, it's easy to forget how relatively small some of them are to AWS. A couple of years ago, AWS was bigger than its next 14 competitors *combined*. A lot have grown since then, but even just a year ago it was still bigger than the next 4 combined:
https://www.parkmycloud.com/bl... [parkmycloud.com]
Oracle doesn't even show up on that list.
Re: (Score:3)
IBM bought RedHat to help them compete in this space.
For 34 billion dollars. ($34,000,000,000.00)
It might be prudent to wait to see what services they roll out before you write them off.
Re: (Score:2)
IBM bought RedHat to help them compete in this space.
For 34 billion dollars. ($34,000,000,000.00)
It might be prudent to wait to see what services they roll out before you write them off.
That $34b sounds like a lot, but in this space it's not really a huge sum - Amazon has already spent far more on AWS and is literally a decade ahead. Heck, $34b is not drastically more than what AWS *made* in 2018 revenue, and every indication is that it will far surpass that in 2019 (see e.g. https://www.statista.com/stati... [statista.com]).
IBM's purchase of RedHat just supports this point - they are far behind and haven't really done any significant innovation in this space, so the RedHat purchase could be seen as tryi
Re: (Score:2)
Conclusion: IBM don't know what CentOS is.
Re: (Score:2)
Perhaps not. Red Hat effectively controls a lot of FOSS software, as proven by the prevalence of Gnome3 and systemD against widespread disgust.
Re: Cringley is a moron (Score:1)
Re: (Score:2)
AWS US-EAST is proven... to have entire datacenter failures multiple times a year.
Re: (Score:3)
And, nobody in their right mind uses Alibaba. May as well just give all your money and IP to the Chinese.
Except the Chinese. There are a lot of them. In fact, the Chinese government could make Alibaba the preferred (or mandated) cloud provider to Chinese companies.
Re: (Score:2)
"And, nobody in their right mind uses Alibaba. May as well just give all your money and IP to the Chinese."
Except for a billion of Chineses, that is.
Re: (Score:3)
and that's only the half of it. Guess what other market they're investing in?
India.
Uh oh looks like they'll be YUGE
Re: (Score:3)
You are a moron, Alibaba has China and taking over India. Guess what little AC tard, that's over a third the human race right there
Re: (Score:2)
Alibaba has China and taking over India. Guess what little AC tard, that's over a third the human race right there
Not all animals are created equal. Specifically, they're not all worth the same amount of profit. Numbers are only meaningful in context. China is likely headed for a crash as economic expansion stalls. India is slow just to get out of the gate. They have money to fix their problems, but won't spend it, so the problems persist.
Re: (Score:2)
No, Microsoft is the biggest by users in the enterprisespace, one corporation in five uses Office 365 and related
Re: (Score:2)
this is the cloud computing thread; you're deranged and fixated, unable to break out of a cognitive obsessive/compulsive loop.
Re: Cringley is a moron (Score:1)
Re: (Score:3)
my employer has the O365 but we host the AD in-house in multiple sites... while I hate Microsoft have to say the Microsoft cloud wares are work better now than the self-hosted stuff. Its adoption will continue to rise, I'll predict it will be there and even bigger a decade from now.
Re: (Score:2)
AliYun Hong Kong has good peering with Chinese and European ISPs. I don't store the master copy of anything important there, but I do use instances there for proxy/cache functionality because using it as a trampoline is often faster than connecting directly to/from stuff in Shanghai from elsewhere.
Re:Cringley is a moron (Score:4, Interesting)
Microsoft is shitty, but they will succeed in the cloud by more or less coercing users of various on-premise products into the cloud.
Even Windows desktop/server users will wind up being forced to use a Microsoft ID to do anything with Windows, and most non-adware versions of Windows will wind up storing user profiles in OneDrive. You'll pay for it and use it whether you want to or not, and most people will wind up using it not be able to weed themselves out of it.
I don't disagree their products in the cloud are pretty stinky now, but people seem drawn to them like moths to a flame.
Re: (Score:2)
Every cloud product they push is a total piece of unusable **** (sharepoint, onedrive, skype for business, etc..)
Isn't this mainly talking about Azure?
Re: (Score:3)
Item one is something he literally apologized for over 20 years ago. Item two is something that's never been adjudicated. Two people claim to be Apple employee #12. Can't find anything that disputes that Mark Stevens (aka Cringley) actually worked there other than another anonymous coward post (from last week).
He has ruffled a lot of feathers and made a lot of bold predictions, many of them wrong.
What will slow the cloud (Score:1)
The EU nations will want to control tax and speech, comments, data, images due to politics and to enforce EU censorship.
China will demand control over its own data due to Communist party censorship.
The USA will offer freedom and try to enter a lot of markets globally only to find local laws d
Nonsense (Score:4, Interesting)
Everything you said is stuff that international corporations have been dealing with since long before the cloud existed.
The prevailing model has always been to hold local corps accountable, regardless of who they are owned by. They often have to modify their offering to comply with local laws. What, exactly, is so special about cloud computing services that makes this less true?
The dominant cloud model is already built on local points of presence. Much of the rest of what you're talking about is a random spray of complaints that some people don't want to comply with local environment. Well sure, they don't, but at the end of the day there's billions of dollars at stake, so they will.
Clearly the greatest advantage for cloud providers is the technical capability to spin up infrastructure, not the physical hosting of it in the United States. For most of the planet, the US is unacceptably distant from a latency perspective anyway.
Re: (Score:2)
In terms of storage limits, quality 24/7 support costs, lower energy costs, better quality encryption, free speech, banking support, bandwidth costs, peering. Legal and tax support. Newer and better servers. Better CPU support for the costs.
The USA was to be supporting the worlds computer network use at a lower cost from the USA.
That all changed withe EU laws, EU censorship, EU nations tax rates, the c
Re: (Score:2)
The cost saving that was to allow the USA to out bid any nations domestic service for cloud support....
That all changed withe EU laws, EU censorship, EU nations tax rates, the censorship demands of a Communist gov in China.
Citation needed. It seems much more likely to me that customers want a data center that's relatively close rather than halfway around the world, for performance reasons. Besides which about half your cost saving factors probably aren't a cost saving compared to other developed economies anyway.
Re: (Score:2)
Hosting in the USA under a US brand has to offer something over services offered in a Germany, France, Japan...
Price, crypto, more CPU power, code support, support staff, tax, freedom, peering, bandwidth?
Re "customers want a data center that's relatively close rather than halfway around the world"
Governments start using laws, censorship and tax to make much better US products and services less attractive?
Protectionism using laws, censorship, tax an
Re: (Score:2)
Hosting in the USA under a US brand has to offer something over services offered in a Germany, France, Japan...
Why?
Re: (Score:2)
US tax and legal protections.
The sending of data to and from the USA is lower cost than trying to buy the same level of support and service in another nation?
Something great makes the global computer community return decade after decade to US products and services over their own nations computer attempts.
Re: (Score:2)
I think either you're mistaken, or I'm misunderstanding you. I believe companies serving their customers in Europe generally choose data centers in Europe, not the US. Similarly for Asia. In addition, there are now serious legal issues with sending EU users' data outside the EU. You seem to be arguing that these companies mostly want to use data centers in the US for various reasons, and I don't think this is correct.
What I know for certain is that the big cloud providers do in fact have data centers all
Re: (Score:2)
Staff costs, CPU costs, better software access. The USA could offer that at a lower cost over time with better service.
It was better and had lower peering costs to use the USA than to pay more to stay in a EU nation.
Thats what set US networking apart from the costs of the EU nations.
Not every service and product needs the fast in nation only networks.
Large amounts of data and CPU cost
Having used Softlayer... (Score:5, Interesting)
AWS will survive and continue to be king of the hill, but it would not surprise me to see it become part of a forced divestiture / anti-trust kind of thing... As a certified M$-Hater, it pains me to say Azure is quite healthy, and works quite well. The interesting things happening between these two right now is the push to move to "serverless" logic. S3 buckets and lambda functions all the way down... Is becoming today's vendor lock in play...
But I must say, having used Softlayer a bit... What the hell are they smoking over at IBM? You go to spin up a template, and it's 20 minutes to two hours before you get anything... Are they paying the janitor to go push buttons on their coffee break!?!? Seriously...
Re: (Score:2)
Really? That long? Instances become available in seconds on Hetzner, and well under five minutes on AliYun. What's IBM doing?
Re: (Score:2)
That's probably why they're rolling out new offerings this year.
Re: (Score:2)
Azure isn't fast either. Recently I tried to convert my 30 GB S1 database to an S2 database. The process timed out after three days.
Too many new players (Score:2)
Oracle is in court claiming Amazon is selling a commodity to the US DOD and therefor shouldn't have an exclusive deal for a long time. The US courts have agreed with that argument in the past. If the courts rule that a specific cloud service is just like any other, then the door is open for all the competition and the big 3 don't have the best price. The 2nd tier cloud providers are much cheaper than the big three and seem to have better tech support.
I've been renting space for my own hardware in data ce
What is AWS' RDB? (Score:2)
> AWS's RDB, in contrast, is based on MySQL and costs Amazon almost nothing to support
Not quite.
If the author meant RDS, that is MariaDB, MySQL, PostgreSQL, MS SQL Server, or Oracle.
If the author meant Amazon Aurora, that's is own technology with an interface that looks like MySQL and one that looks like PostgreSQL but it is neither one.
3.5 (Score:2)
What's this Cloud v3.5?
I thought bob wasn't doing these anymore. (Score:2)
I stopped reading his predictions column after he said he was going to stop doing them, five years running. Glad to see he's still at it though, but I really feel like I've moved on.
IBM and Oracle merge (Score:1)
Is it possible IBM and Oracle merge to gain cloud muscle?
China (Score:1)
AMZ business tactics are no different than Chinese;
I am no Cringely (Score:2)
but I predict a switch to private cloud services (e.g. cloud service running on a set of hardware and software belonging to a corporation) in about 5-10 years. whoever comes up with a best offer, will dominate the field for the following 20 years or so.
Oracle and MS have strengths; IBM has none (Score:3)
MySQL is also available in the Oracle cloud: https://cloud.oracle.com/mysql [oracle.com]
Of course it's the Enterprise version and I'm sure that internally Oracle (mysql division) are billing oracle (cloud division) in funny-money, but the actual cost to provide it for Oracle is very low. There aren't many differences, or much coding effort, between the Community (Free) and Enterprise MySQL - just enough to get Enterprises to pay for the extra audit and thread pool (i.e. helpers for crappy applications that can't use a database correctly) support.
However the big point of Oracle cloud is not that it has MySQL, but that it can supply your on-premises private cloud infrastructure as well as off-premises public cloud. Their aim is to satisfy those people who want their data on-premises for one reason or another, but don't want to have to do the work of building that infrastructure themselves.
MS Azure cloud is wildly popular with Linux people, rather to MS's surprise - there are far more Linux customers than Windows Server customers in Azure. Meanwhile, if you want to run your desktop app back end in the cloud (i.e. Office 365) and have decent Windows hosting, Azure will do that for you with one supplier contract. That's a really strong advantage of Azure. Microsoft still has the global hold on office applications and they can, if they're reasonably smart, transition that into becoming the "inevitable" cloud supplier for companies with a lot of office-application users.
IBM... isn't looking like it has any of the advantages. They don't have the advantage of being the first choice for Chinese companies, nor the cheapest and biggest, nor the public/private single interface, nor the obvious place to keep your MS desktop apps while closing your datacenter. They're also late to the game.
Re: (Score:3)
The big problem with Oracle's cloud is that it's being run by Oracle. The distinguishing feature of Oracle is that they will always try to fuck you over. Sure, everyone does it sometimes, but Oracle does it always.
Re: (Score:2)
Oracle doesn't always, they usually wait until they can inflict maximum damage
I wonder (Score:2)
...who the .5 one will be.
There are cloud providers everywhere (Score:2)
MS, Amazon, IBM, Google, they are just the "big" players. My company uses a much smaller provider called Omnipotech. These smaller providers are everywhere, and they are not going away just because the big boys are fighting for market share.
RDB? (Score:2)
Lots of pressure will come to bear in this case from IBM, Microsoft, and Oracle, who are all suffering from a very specific database problem competing with AWS. Each of these companies sells their own database (DB2, SQL Server, and Oracle, respectively) that they've rolled into their cloud services. AWS's RDB, in contrast, is based on MySQL and costs Amazon almost nothing to support, giving the biggest cloud player a clear pricing advantage.
This is not true.
There are a wide range of database engines to choose in AWS. RDS [amazon.com] explicitly lets you choose from Postgres, MySQL, MariaDB or - Oracle or even SQL Server. And guess what else? Microsoft themselves offer a managed MariaDB instance [microsoft.com] on Azure! This guys post is straight-up bullshit and looks like he did zero fact-checking.
We're still listening to this guy? (Score:1)
Re: (Score:1)