Donnacha MacGloinn Predicts the End of the Webhosting Industry

Donnacha MacGloinn published an excellent post on predicting the end of the webhosting industry. Although he talks about the industry as a whole, MacGloinn says the need for specialized WordPress hosting is quickly disappearing.

With the lingering perception that this stuff must be difficult, and the customer’s delight with all that WordPress can do these days, the specialist WordPress hosts get a lot more credit than they deserve.

The problem is that modern technologies are rapidly rendering their role unnecessary. Anyone can now go to Linode or Digital Ocean and fire up their own VPS (Virtual Private Server) within half a minute, for a fraction of the approximately $30 cost of the most basic, one-WordPress-installation package from a specialist WordPress host.

The VPS user can then use Docker or a similar technology to instantly install all they need for any number of secure, optimized installations of WordPress, all running with more memory, storage and bandwidth. The dirty secret of the specialist WordPress hosting industry: this is exactly what most of them are doing themselves.

If specialized WordPress hosts rely on cloud infrastructure services like Amazon Web Services and charge a premium to customers, what stops Amazon from providing a great managed WordPress hosting experience, thereby eliminating the middleman? As MacGloinn points out, Amazon, Google, and Microsoft have already made moves to provide the best out-of-the box experience.

Just last Friday, Google announced Google Cloud Launcher, enabling push-button deployment of over 120 top Open Source applications, including WordPress, to the Google Cloud Platform. Just two days before that, Microsoft launched the Azure App Service, which does essentially the same thing from a slightly different angle.

VPS (Virtual Private Servers) and other technical jargon is scary to a lot of people, but if companies like Amazon, Google, and Microsoft can provide a great out-of-the box experience and charge a significantly lower price compared to specialized hosts, it’s possible the managed WordPress hosting space could disappear.

The article has a few comments from readers and offers a lot of food for thought, but I really want to hear from those who own and operate specialized WordPress hosting companies. Do you think MacGloinn is on to something, or is his prediction way off base?


55 responses to “Donnacha MacGloinn Predicts the End of the Webhosting Industry”

  1. MacGloinns opinion sounds bit like the “Manifest” by Karl Marx: basically true – but there is some other reality that does not fit into todays picture – maybe some days later. Users / Customers of the web all will not rent a core server and operate with docker or whatever. The next 10 years are about the content, not the technic how to provide this. And todays tehnic is far far away from to be managed by content providers.

    • Hi Andreas, thanks for the comment.

      With regard to the Karl Marx comparison, I would like to assure everyone that I shaved just an hour ago. My communist tendencies go way down when I am clean-shaven.

      My article does not address the issue of content which, I agree, is the most important part of any website. What I am looking at is the possible effect of emerging server technologies upon the many different forms of Web hosting we buy. The initial effect has been to make it far easier and cheaper for those providers to offer stable, reliable and performant hosting – that is why there has been such an explosion in the number of specialized WordPress hosts in the past few years.

      I am predicting, however, that the longer term affect of all this automatation will be to make it far easier for the cloud giants to enter the consumer hosting market with credible offerings, and I suspect that WordPress hosting could actually be their first target.

  2. Although it will likely make a dent in the hosting service provider industry, I do not think it will lead the their demise entirely. There are always people who would rather pay someone to do things for them – for a variety of different reasons. Outsourcing anything depends on time, effort and money. Those who are willing to pay someone else will continue doing so. Just because I know how to paint, doesn’t mean that I should always be the one painting…

    • Hi Toni, thanks for commenting.

      Sure, this will all take time, inertia is built into anything we humans do. I spread my prediction out across a decade, with a bonus peak at the distressing nightmare world that awaits us in two decades.

      With regard to paying other people to do things, yes, but when it becomes apparent that a particular bundled service is not providing value, the market tends to reconfigure. Today, the hosting industry bundles together the “metal” (the actual servers, bandwidth etc) and the support.

      When you buy, much of the emphasis is on the support aspect, because the hosting companies know that the vast majority of customers are clueless but afraid. They also massively overpromise, allowing their marketing text to suggest that they will happily deal with any problems that arise, while their small-print reveals the extremely limited circumstances in which they will apply actual support time without additional charges.

      One of my guesses is that, although a substantial portion of the hosting industry will trundle on exactly as they always have, a lot of customers will defect and most of the new growth will be soaked up by the cloud giants offer of fast, reliable, stable, fully-featured and cheap hosting. If they can promise users that they will keep their WordPress sites, and a white list of plugins, continually updated and blazingly fast, that, along with constant monitoring, backing up etc will be enough for 90% of WordPress users.

  3. In order for me to put any real trust into what you’re saying. I’m going to need you to provide more information about backups? I want to know more about security? I want to know more about optimization? If you’re not going to be able to secure, back up. To firewall protection along with actual word press knowledge. This is where you fail to understand paying a premium for something. I do not wantyour fake firewall thank you I’ll stick to Pagely’s VPS & my FireHost server.

    It all comes down to how much time do you have to spend on something if you’re telling me that is something you

    • Hi Thomas, thanks for commenting.

      Actually, all the features you mentioned are actually the ones that are easiest to automate and, therefore, are entirely suited to how the cloud giants (Amazon, Google, Microsoft) operate. There is no such thing as a “fake firewall” :D

      The sticking point is always going to be support: to what extent will they be willing to break with their usual formulas and provide their hosting customers with real, human support? We already have good indicators, from existing products such as Google Apps for Work, of how they will approach that.

      • Donnacha,
        Great post, this is very timely for me. Most of my work has been on CPanel types of hosts in the past. I was just asked to work on a site on one of these “super duper WP-specific” hosts. With just a bit of sleuthing I discovered that it was on Linode, so it was exactly as you said.

        The experience was mostly OK. They have a staging area, which is handy, even though there are odd little things, such as uploaded files that don’t load in the staging site even though I can see them in the file system and permissions were good. And the truth is, I can get by with the limited number of tools they offer because I’m very self-sufficient. Interestingly, I found that performance was good, but just as many slow times as regular hosts. I suppose it’s possible that staging gets less juice than production.

        I definitely agree with you that they will grow and threaten non-cloud hosting companies, now that cloud hosting is more reliable. It’s true that support will be a key factor, although I usually don’t find myself using it my travels.

        They will be sufficient for mildly technically skilled users who don’t want much customization. But those who love to tinker beyond their enormous WP theme admin panels will still need to hire help, just as before.

        Thanks for writing this!

        • Dave, thanks for the kind words of encouragement. With the regard to the limited nature of the support that is actually available, you are absolutely right – anyone who is sufficiently curious about WordPress to read websites such as WPTavern will very rapidly outgrow the level of support that any of the specialist hosts can provide to them. What they are really selling is reassurance for all the people who have heard, in the past, that WordPress is somehow problematic or insecure.

          I can guess the host you are talking about and, yes, they are actually repeatedly, consistently, horrifyingly bad. I have been hired in to give advice in situations where big clients, paying huge amounts per month, were being messed around for months over what were actually extremely straightforward, obvious and easy-to-solve technical problems. I won’t name the company but I will give this tip to anyone considering buying the services of any specialist WordPress host: take a very close look at the company structure and ask yourself, are these guys taking the time to grow organically and ensure that all their staff have time to learn, or are they pumping millions of investment dollars into rapid expansion?

          • Donnacha,
            Nudge nudge, wink wink, as to the host. ;) I guess I mostly lucked out while there, oddly enough.

            To your good point about newbie reassurance, WP is, to some degree, a victim of its own success. You have newish people who have never used anything else, so they are in that bubble. They’re easy marks for “WP” hosting. And being in the bubble, they stay unaware of the big threat of alternatives to WP itself.

            Unlike the rest of the social media world, where over-sharing is ubiquitous, you’re right, hosting businesses (and many other WP-oriented businesses) are very opaque about their policies, activities, and so on.

            My own host doesn’t mention WP at all in its listed offerings. They are just really good, give great support, and don’t oversell their servers.

            Best, Dave

  4. Let’s face it, web hosting will be free soon, free as in Facebook.

    It won’t cost you a nickel in exchange for indexing your content (somehow deeper than they do now), site traffic demographics and that of your audience/visitors.

    This data will help “optimize the user experience.”

    • Yeah it’ll be free like datacenter space, power, bandwidth, server hardware, skilled labour. All for free. They’ll give it away because non of that costs any money at all.

      • Re-read my comment, Some Dude.

        The end user is paying with their data. Welcome to the internet.

    • FREE is certainly an impressive price-point, but it distracts from the real revolution: the increasing range of tasks that can be automated or performed by a “normal” customer via an improved interface. A few dollars does not matter anywhere near as much as the democratizing trend of previously “expert” skills becoming someone you can do for yourself.

      • But that revolution has been happening for over 20 years. I think the piece you’re missing is innovation. Just because you think in 10-20 years what we have today will be automated, doesn’t mean the new technologies built in those 10-20 years that everyone wants will be automated too. When everyone can have the ‘expert’ skills of today, the experts have moved on to tomorrow. Look at things like DreamWeaver. It automated the building of HTML sites and let users create websites without coding. That didn’t kill the market for web developers. Web developers made cooler shit instead.

  5. I’m not too familiar with VPS servers, and I have a question about this. I know this isn’t the point of the article, but this seems like the right place to find the answer:

    So if I’m a web developer / consultant who normally hosts client websites on a web hosting account, would it be better / more economical for me to sign up for Linode or Digital Ocean and get a VPS? Does that require more technical know-how to set up? Sorry if some of this is over my head a little, but I do want to understand it.

    • Good question Paul. There is so much disinformation out there about this, precisely because so many jobs depend on keeping this stuff mysterious.

      Simply clear 3 hours in your schedule and find out for yourself by firing up a Linode and applying a Webuzo stack. That will give you an interface very similar to the CPanel interface that pretty much all Web hosts provide.

      If you have some spare domains to hand, experiment by creating some temporary WordPress sites, and be sure to look at all the other applications that you can install at the touch of a button.

      Then, have a thorough look at all the different settings on your Webuzo end-user panel and ask yourself: “Is there anything I’ve asked my Web hosting support to do for me, in the past year, that I could not have done myself from here?”. Even the relatively arcane stuff, such as changing PHP settings, is right there.

      If you have paying clients, money is not the main concern, at least not directly. Obviously, being able to buy more resources – especially more RAM – for your money will prevent some problems from ever even happening, but the main advantage is that, by absorbing this particular set of skills, you can move a lot faster and be far more responsive to your clients’ needs. You will be able to make changes far faster than it would have taken you to write a ticket asking someone else to do it.

      This is just a relatively simple example of how people can jump up the skill-chain, even with existing technologies, my prediction is that it will continue to get easier.

      • Also of note for those who do want deal with best performance be it WordPress or Jooml’r.

        When you look at the costs of many VPS’s or shared hostings along with VPS etc. you rapidly will see that dedicated hosting is actually a better choice. In respect to resource usage slap as much as possible into SSD drive(s) on the server.

        Not only does this vastly increase “Peppy” nature since files come off a solid state device .vs a mechanical drive but by increasing that throughput results in your resources being released faster.

        • Today, dedicated servers will often give you better value in terms of raw resources, but what I am focusing on is the way in which the far more granular level of VPS hosting allows us to do radically different things. Looking years ahead, I am not particularly worried about cost because that will continue to become a less and less significant part of the equation, I am more interested in how ever-improving server tools allow us to push human involvement (by far the most expensive cost) further up the value chain i.e. allow intelligent people to work on site design and content, rather than worry about site deployment etc.

      • donnacha, I really like your answer here! I’ve have set out to find out how to work on a VPS once, sat down with a few tutorials and couldn’t figure out a thing. :( I’d consider myself a pretty capable guy as far as getting around in a cPanel or two.

        What I need to do is watch someone do it once. YouTube has NO effective videos of someone setting up WordPress on a VPS. I think the first company to screencast the setup will be the first big WP + VPS solution for the masses! Any takers out there?

        • That’s a good idea rgregory, perhaps I’ll try to make a few videos, the main problem is that this area is moving so fast, videos won’t remain accurate for long.

      • Your article makes some good points and maybe Google, Amazon, Digital Ocean, and friends are starting to ‘democratize’ hosting the way WordPress is democratizing web publishing. However, your answer to Paul seems incomplete.

        While I think that Linode is a great platform and if one is interested I think that ‘diving in’ is a good way to experiment and learn, it seems like you are minimizing some of the dimensions of being a server admin, which is the role you take on when you administer a Linode VPS.

        Installing a linux image and Webuzo is only part of the picture. Linode backups (not site backups) are an extra cost and Webuzo for WordPress is not free. Webuzo does not provide an interface for OS updates and administration. You are running a server on the Internet. Are you confident that you have setup IP Tables, Fail2Ban, TripWire, SSL cert etc correctly? There are guides for these things, but it is not a 3 hour task, it is journey that can take months or longer to master.

        If those things are something that you want to do then you will feel empowered and it a joy. If you just want to blog or focus on your website then those things are a pain.

        • Hi David, thanks for the comment. Any guide to running a server, squeezed within the confines of a blog comment, is doomed to be incomplete, but Paul was asking for advice on how to get started. I believe I pointed him in good direction, one that will enable to overcome the barrier that prevents most people from even investigating the possibilities: fear.

          I only recommended Linode because they provide a push-button Webuzo installation, and I only suggested Webuzo because that makes pretty much everything else push-button. I said 3 hours because that will be enough to give him an overview of what might be involved. Webuzo has a free trial and Linode for those 3 hours will cost him five cents … he can think of it as an investment.

          At the end of those 3 hours, he will have something to contrast against his previous experience with CPanel. He will either decide that he does not want to take on the extra complexity or he will be hooked by the realisation that his work can potentially be faster and more effective if he takes control of this layer of the stack.

          That will be sufficient incentive for his to spend a few more evenings running through the many excellent tutorials on securing your server, server-level caching etc, allowing him to move confidently forward from both Linode and Webuzo. I don’t accept that learning how to competently manage a server would take more than a week of 3-hour evening sessions.

          Incidentally, you asked the question “Are you confident that you have setup IP Tables, Fail2Ban, TripWire, SSL cert etc correctly” … let me flip on that its head and ask you: why are you so sure that the low-wage, inexperienced kid hired by your hosting company knows what he is doing? I would argue that, following tutorials and, of course, running the included tests which confirm you have completed the steps correctly, allows you to be far more confident.

          Incidentally, if you think I am joking or exaggerating the possibility of hosting company staff messing up, I am not. The bulk of my earnings over the past decade have come from mopping up after the spectacularly bone-headed actions of hosts, including the supposed WordPress experts, and they are steadily getting worse, not better.

  6. Remember time, when building websites was for programmer, technical people. Now almost everyone can build nice website using WordPress or anything similar.

    Something like that will probably replace old fashioned hosting service soon … just tool like ServerPilot, few clicks and everything is set up on cloud without technical knowledge.

    • Peter, very good point about WordPress which is, itself, just one example of the sort of gradual software improvement I’m talking about, which expands empowerment to an evermore mainstream audience – what we see as “expert” skills today may become rather rudimentary tomorrow.

    • Wix, etc. are entities in the infancy of all of this. The web in scant years will be drag, drop, associate data and done. More like a combination of a paint program and word processor. Sorta Desktop Publishing ++ online.

      That is what Adobe is working towards for example.

      I forgot what my brothers buddy called it, he works for em’. Multi something or another interactive community publishing something or another. They are as one might guess WAY ahead of the rest since that stuff’s been their gig for years in as far as the codebase goes I am told.

      They have not went to their cloud based applications because “Well we need the word cloud too”. Its all about this interactive distributed publishing platform.

      According to what my brother told me it contains three major gigs for web folk. 1. Its completely scalable from word go. 2. It can be easily used by anyone who can use something like Printshop on their PC or be as complex as General Motors might need to put forth an international presence in all nations specific to those nations. 3. It affords a complete and easily alterable workflow so if GM US, GM Europe, GM Japan all want exact workflows and data flows they can have that. But In China they need it different, they can do that.

      Lastly, possibly as soon as next year it will be available.

      • I think all the companies in the general space know the direction in which things are going and are anxious not to get locked out. Both Adobe and Microsoft are working on impressive stuff now, precisely because they fumbled their previously dominant positions so badly. All of this is terrific news for anyone who works with the Web.

  7. So, not so much “the End of the Webhosting Industry” as just the end of specialty hosts that only support one specific platform.

    • Hi James. No, my article is very clear: this disruption is barreling towards the hosting industry as a whole, I’m just predicting that it will become apparent more quickly in the specialty WordPress hosting niche.

      • Well, the conclusion of your article entitled “The End of The Web Hosting Industry” seems to be a bit of stretch to me if your conclusion is that “within a decade from now we are likely to see cutting-edge hosting, with deeply integrated applications, provided as part of Amazon Prime or Google accounts.”

        It seems that you’re suggesting that new companies will enter the web hosting industry, pushing improvement and innovation, and those who do not will fade away. It’s not the end of the web hosting industry, just a pivot or evolution of offerings.

        If the web hosting industry were to end, there would be no hosting. Your theory suggests there will instead be *better* hosting.

        Sorry, I guess I was just put off by the title.

        • James, you’re attributing meanings to my article that just aren’t there.

          Far from focusing on the entry of new companies, I am saying that the existing cloud giants, who have been around for a looooong time (Microsoft 1975, Amazon 1994 and Google 1998) are already gearing up to swoop down upon the existing industry and all but wipe it out as a standalone entity.

          Today, it is still possible to buy both horses and horse carriages, but the horse carriage industry can fairly be said to have been wiped out by another form of transport, the automobile. There is still transport, you can still climb into a vehicle, but the industry that had grown up around the provision of horse-drawn transport was annihilated. I am saying that the same thing will happen to the industry that has grown up around hosting. There will still be hosting, but it will be provided by very different companies and on a very different basis.

          Don’t worry, your job is probably safe: as I say in my article, it is more than likely that companies such and Automattic and Square will be purchased as the cloud giants fight to gain an edge (Pro-tip: if Automattic gets bought by Microsoft, don’t let them give you Ballmer’s old office – none of the chairs in there look right).

          • Donnacha, you’re attributing meanings to my comments that just aren’t there. :)

            I’m not worried about the future of hosting, nor am I worried about the future of my job. I just took issue with the “The End of The Web Hosting Industry” headline on an article that was more about its evolution than its end, to me at least.

  8. He is actually 100% correct but misses much in the points made. The Gobbling up of the enterprise Internet already has for the most part saw its days. The next logical foray for corporations to Gobble up becomes the next largest segments or demographics of the Net. This does not only include specialized hosting provision. We use Codero and quite frankly I’d put them up against any “Specialized” hosting firm for Magento, WordPress or any other and can pretty much assure you they come out on top. Its dedicated hosting but for the price smashes any VPS form of provisions, its not even comparable.

    All that aside. Cloud based enterprise is what brings forward smart device unification. Its core to it and the speed of this unification is of utmost priority, not necessarily to business but to governance of nations and regulatory measures over this communications medium.

    That is a primary goal right now as we speak globally. GeoIP does not exist for sake of one being able to put “Welcome from Seattle, Wa.” on a WordPress site. Nor does it exist to triangulate cellular towers to maximize bandwidth, nor does it exist for GPS systems even though all those and many more take advantage of what GeoIP can do. Its main reason for coming into existence is to regulate. Thats global governments, not business even though it affords business opportunity and better routing/control over infrastructure. Its about everything from nations security, privacy (or lack thereof) and being able to regulate.

    The “open source” success stories are also in the gunsights of business. WHY would they NOT be? Really. WordPress states that some 20%+ of the net runs WordPress sites. That INSTANTLY makes it a target for corporate business. It is coded in PHP which puts it NOW due to .NET’s increases in performance and considerably more a target. We all do know? Yes? Microsoft, Adobe, Apple, Amazon, Google etc etc are aggressive corporations yes? Does ANYONE honestly think they are just going to sit on their duff’s while in WordPress case 20%+ of websites run it?

    Not hardly. They will and in fact are going to introduce new software. Software that is WAY beyond the capabilities of WordPress while being “state of the art”. But ALSO at the sametime under the operating environment that THEY control. And then, regulatory measures can be “real” THROUGH the software base operating environments INSTEAD of thinking webmasters will simply comply.

    This is WHY I stated MANY times the WordPress core developers OUGHT NOT be too all concerned with where WordPress is right now. Set a few people to still work on it. But the CORE group best be learning C# if they dont know it and working feverishly in fact to create WordPress II that while perhaps resembles the current WordPress (and can convert an existing site with a click) it needs to vehemently stick to ASP.NET / C# and API code standards as if it doesnt, it wont sit well in the unification of smart technologies. Anything other than that, its just sadly dead and just doesnt know it yet.

    I know of two very popular Open Source shopping cart applications that are already doing this. They do not have options. Its coming to a “Cant beat them so better join them” and again, the rate is becoming more feverish if ya’ll have not noticed. Enterprise entities taking significant leaps that have in reality been being worked on unbeknownst to 99% of us for YEARS.

    A new “shakeout” is on the horizon. Consumers will embrace it completely as it will allow them unparalleled capabilities under some moniker “Web 5.0” and in time remove freedoms via cause and effect events.

    I just went through some of this with a friend in fact. Horrified by the shootings today in Kenya. Why is religious clash in the form of terror on the rise? While never being a peaceful per se coexistence throughout history our times see such drastic distress? It is in large part due to cultural clash. Old power .vs. new culture. Clerics loosing authority in the wake of populations, nations, accepting new things and with those new things come new ways and usually loss of old ways.

    It will be and is truly ugly.

    The Internet is part of that and the only way it changes is for those who create the technology to change it and that my friends is on the horizon.

    Not only will you see alot of Gobbling up of hosts, you will see the Gobbling up of so much more.

  9. IMO, I think this article was written looking for some internet attention.

    The model for having a bunch of specialty images just isn’t viable. I mean, let’s break down just the WordPress specialty industry. If you go to a group like Advance WordPress on FaceBook and scroll through, you’ll see question after question about server configurations. Some people are pushing hhvm, other varnish, and more recent php versions, and nginx, and nginx in front of apache, or even better, varnish as a caching service with nginx as a front-end proxy to apache.

    My point is there will be so many different configuration that people want that it won’t be possible to create drag-n-drop images for all of them. And even when someone finds an image they want, there will inevitably be some little tweak they want made.

    And let’s look at another thing. Companies want to pay as little as possible and make the most money out of it. Companies aren’t going to pay WordPress experts, Drupal experts, Joomla! experts, and any other platform experts to sit down and figure out all these images and get them finely tuned. Because they won’t pay for it, the services will be lower quality, and we quickly get into the “shared hosting” situation people are always complaining about.

    At some point, “developers” need to own up to the fact that they write shoddy code.

    This just doesn’t make sense.

  10. “VPS (Virtual Private Servers) and other technical jargon is scary to a lot of people, but if companies like Amazon, Google, and Microsoft can provide a great out-of-the box experience and charge a significantly lower price compared to specialized hosts, it’s possible the managed WordPress hosting space could disappear.”
    I think the article is correct in the sense that the layer of “middlemen” will become smaller. However, there will always remain a layer I believe. Most people who want a website build don’t even know what a “server” or “VPS” is. They don’t know the difference between Windows or Linux servers. Etc etc. Pointing them to the AWS or Google Cloud page and trying to have them understand what is explained there, no chance of succeeding. Even for me as a web developer it’s not all that clear, there are so many options. And sure, I could spend a day learning about it, but if I’m building a website for a client, having them signup for any good shared hosting package (with one of those middlemen!) in just a few clicks for $50/yr is just as easy. $50-100/yr is so low already, that there’s hardly an incentive to try to reduce that even further. If a businessman/women needs to spend an hour studying the options on Google Cloud, AWS, Microsoft cloud, etc, he already wasted more money then he can save.
    So maybe the future web hosting service providers will stop running their own servers and use the big cloud platforms more and more. But that middlemen service between the technique (managing the server) and the non-tech people who want a website/app build, will remain.

    • Many have already done this and more are doing so. Completely transparent to we poor peep’s.

      My brother before he left the #1 dating service where he was a lead engineer did exactly that. He and his group moving the entire codebase and data to Amazons cloud completely eliminating the expense of the hardware infrastructure which costs a TON more to maintain both in hardware and human resources than the Cloud.

  11. I’ve used a bunch of as you say “shoddy” VPS arrangements and of course the same it true of shared hosts. This is why I finally had enough, went to Codero and have a dedicated host. In the long run it costs less money and the fact I have SSD’s setup and altered code to have most common files used for say WP installations grabbing them off the SSD makes for more of a performance increase that all the tweaks one wish do with Apache or NGinx. Why fart around of you will instead of just doing it right in the first place.

    Shoddy code is another matter completely. One developers thoughts of shoddy code are not anothers. If you care read any of the folks who have actually dealt with code performance in historical context such as Abrash you will find that the non-traditional fashions of looking at structured problems solved in non structured ways often provide the best performance at the cost of code maintainability. Years back I was in that movement. I and my brother authored ST Lightning a complete replacement graphics kernel for Atari and Amiga PC’s. Hand tuned assembler code .vs. what was there which was compiled C. Performance increases in some cases were just off the charts. For example, the second most used primitive, line drawing (diagonal) was traditionally done DDL. Much like on a graph paper. Step 1 pixel this way, 2 that way etc. tracking the actual line segment error. When the error grew more than 1 pixel (as Raster displayscould draw a perfect diagonal line) the error was corrected by a plot, reset the tracking axis and continue on. We instead plotted to RAM (vs display ram) until that same error occurred but then instead of correcting and plotting on we corrected then replicated the previous plotted line segment and do a Bit Block Transfer into Video Ram. WHAM. Increases on the magnitude of thousands of percentiles in many cases. Its first use was in Quake as they had problems punching out decent frame rates. This when the WWW didnt even exist, was UseNet etc.

    To the Math whiz weary what we did was just WRONG to do, but to entertainment (games) and imaging applications it was big kicks. I worked with Atari, EA, Sierra, EMI to name a few.

    These days all has changed sorta. Code maintainability is more a paradigm than efficiency and of course applications are often more complex at least on PC’s and Mac’s. On the net, really is sketchy. Donkey Kong is a more advanced application to code that is WordPress or Joomla. A video game such as Halo 5 in engineering terms is just way beyond.

    On the Net the game has been due to the nature of finite resources (ram, CPU cycles) and it sorta (yet not) being multi-user to have small scripts do chunks of work. The Internet is also a stateless.

    PHP came up as the winner in scripting for many reasons. Most notably the complexities of using C++ for socket based work along with some other matters. If one looks at the few C++ frameworks that exist for web use they are considerably faster than anything in the PHP world and Java took the wind out of what progress was being made with C++ Java code being portable through the JVM execution engine.

    Java will happily run under Linux, Irix, Windows Server and pretty much every other environment the JVM is moved to.

    Microsoft suffered unto Windows Server. The environment of ASP.NET, C# (MSIL), Windows Server and MS SQL server is a VERY large stack. It IS a more capable stack in that things like permissions are completely built into .NET. Your PC is more capable for example to run applications than through a browser over a client/server w/ HTML or XML markup for the UI interaction. Those same Windows applications are built atop .NET as is Windows.

    Its a much larger environment but not simply due to say bloat, its simply put more capable. Enterprise ready “Out of the box”, “Distributed application ready” out of the box etc.

    Oracles largest competitor in enterprise database is Microsoft. They did not buy out mySQL for sake of concern against Oracle’s offerings. They bought it so they have a hedge in the marketplace that they can solely control.

    All that aside….

    With PHP it was NEVER meant to become a general purpose language, it was a pre-processor just like C++ has a pre-processor. Writing Object Oriented code in PHP results in significant performance hits and even more significant resource hits. While I have not toyed about with the latest greatest past few releases of PHP I can say a year and a half back or so instantiating a simple class took near 1K byte without doing a thing. Scant few setter/getter methods and a simple class.

    Single threaded still while virtually every server out there is running multicore CPU’s. VPS arguably exists due to the lack of threading ability in PHP. Hosts wanting ways to maximize their hardware.

    What ARE the choices? Java is beyond the avg person wanting run some webs. The average PHP coder hasnt embraced Java yet Java is superior in performance, scalability, resource usage and much more. But, it is a full blown applications language. C# or Visual Basic on the other hand are VERY intuitive to use. Great pains have been made to get it there. Yes, its still far more complex than PHP. One can code mobile app’s, desktop app’s, web app’s, console video games, enterprise app’s, distributed enterprise app’s, client/server app’s and more. But now they (Microsoft) have decided to target the Open Source market. Thats what Mono is. Its why .NET source code is now available.

    As I said, two of the most popular Open Source PHP shopping cart applications are now in code moving to ASP.NET. That is the the future. There is simply no possible way the “PHP Camp” developers are going to “re-invent” PHP to compete against the .NET platform. They do not have the resources to do it, not in programmers, not in time, not in cash. And even if they could do so all those things aside the changes would break all existing PHP code.

    Where languages such as Java, C++, C#, VB encourage coding standards by nature of the language with strong typing, defined interfaces to libraries this is not the case with PHP.

    Those that like argue of PHP .vs. Java .vs. C# (ASP.NET) more often than not have never used Java or C# in any significant manner. Everyone wants to think what they do is the best. I am not like that. I have been coding since I was 17, thats 36 years. I started with Assembler on the 6502 and 8080 CPU’s. CP/M and TRS-DOS. FAST is assembler, least resource usage is assembler. If you took a WordPress script, one of the functions. Coded it in assembler. Ran a benchmark between the PHP .vs. Assembler version you would find the Assembler version on the magnitude of thousands of times faster in most cases.

    We coded three Apache Extensions for a project. These addon’s were based towards Joomla which we wanted to use as a CMS managing those portions of a much larger project.

    Joomla in unit tests could not handle the expectant server load. When we slapped a low level profiler handed to us by Xerox corp (local to us) on the server we saw: Enormous performance issues in PHP (presumably why Facebook wrote their compiler), we saw the same thing. We saw HUGE hits in database connection times, I mean HUGE, the concept of a singleton being good for database access is just plain stupid especially when the database (mySQL) is not only capable of but in fact better AT multiple connections access and performance. It wasnt built for PHP.

    Templating, Seperation of display logic and models. More big hits (Smarty).

    So we coded a smart cache in assembler. We coded a database pooling extension that uses both C++ and inline assembler and a template engine also part C++ and part assembler. We make all these alterations to the Joomla core files to use our extensions. Bam, first run was over 400% increase in performance. In continuing to work with it we coded a easy peasy data mining converter and mined 3000 pages from Wikipedia as test data and used some Rocket Theme heavy weight template. We (actually a Xerox programmer) created an intermediate cache that sat between ours, the DB pooling and templater. Instantly, 750% performance increase.

    For giggles one day we slapped a LAMP stack on a Pentium 4 box and one on our internal Core i5 server. We installed all our altered code on the P4 and a standard J! install on the i5. We imported the Database. Guess what? The P4 was faster in work in -> work out. We used automation software provided by HP to run the unit tests.

    BUT! Even at 750% we are not at expectant traffic loading.

    The CORE application is all being done in C# now. The CMS we are looking at WordPress which is why I am even here.

    Now sifting through the core WP code to see what we may need do IF we do it. What we need do for the Smart Cache, DB Pooling and getting the separation of the theme related matters using our templating engine as the engine itself is faster than straight PHP outputs. It runs on its own thread (can use 4). One call made to it for session setup, from there all output writes go to it and it performs all interaction with Apache. The code the coder at Xerox wrote basically interfaces all three modules to use the smart cache “intelligently” based on session demands.

    When it comes to “bad code” doesnt matter the language. I can write code in assembler that is slow by intent.

    But over the two year journey in this application we are making there are things that stand out.
    Two years ago PHP was a better choice than .NET / C#. It no longer is and will continue to wane because for it not to would require enormous changes in its code base that would break literally all PHP applications.

    That Object Oriented Code in PHP if one is seeking best performance and minimal resource usage should be steered away from in favor of linear code and Arrays. The lack of effective multiple thread support is a severe setback not that it would matter as applications would need significant changes to their codebases to incorporate it breaking most “modules, components, plugins whatall in the process”. Make SURE variable typing is enforced in your code thus PHP not needing do type conversion constantly. Set up your own mySQL tables where applications instead want serialize and unserialize for data store essentially destroying a basic concept of relational Databases, normalization.

    RUN your PHP applications off SSD drives or preferably the entire LAMP stack.

    mySQL works wonderfully with .NET as well as does sqlLite etc. For whatever the reasons PHP coders continually think if they are working with the Microsoft environment they are bound to everything Microsoft and thats the farthest thing from the truth. There are actually more resources, hardened, tested in enterprise available for .NET than is the case with PHP. Literally anything one can think of in fact.

  12. Well that was a bunch of useless information… “shoddy” code is in reference to code efficiency. If you write code that is prone to memory leaks, you write shoddy code. If you write code that has too many, especially inefficient, sql queries, you write shoddy code. If you write any code that has a negative impact on the server such that you suffer in experience, you write shoddy code.

    And of course, I’m keeping this in the WordPress context. WordPress will never move to Java, .NET, C# no matter how compatible they are.

  13. Thats funny. Obviously you have never worked in any enterprise environments. Shoddy Database 101 literally as is taught in first year college is breaking rules of normalization in a RDBMS. WordPress context? Exactly how much of the core source code have you studied?

    WordPress was already ported to .NET once and the results were 300-400% performance curve increase “out of the box”.

    Dont say never.

    There was talk here about it in fact at a “Word Camp” in Buffalo not to long back after the yip yap talk. I dont know if anything ever happened. They were talking about creating a fork of WP to either .NET or Java and reselling it. Some company from Korea or China. I never followed up on it.

  14. To me it sounds like you are predicting the consolidation of web hosts to just a few large cloud hosting companies more than the “end of web hosting”. Cloud hosts are web hosts. The tendency you are describing is not new and has been expected for a long time. Windows Azure for example started shipping one-click WordPress installs two years ago.

    What concerns me is toe monopolization of storage and processing this may lead to, especially because most of these services are US based.

  15. It’s not fair to dismiss all WordPress hosts by saying they don’t actually provide the 24/7 support they claim to, or aren’t actually experts. That’s the primary reason I use WP Engine. They help right away when something goes wrong. Not with just a response, but with actual help moving me towards a solution. And yes, they are technically competent.

    If my site goes down for 24 hours, let alone a few days, because I’m waiting for help or can’t fix it myself, I’ve more than lost the money I would have spent on a month of hosting.

    If Google provides cheap hosting and a slick interface that anyone can use, that would be awesome, but there will always be a place for premium services. Some webmasters, myself included, want to spend their time doing what they do best, and let someone else manage the hosting, backups, security, performance, etc.

    Not everyone is a bozo wasting their money like you make it seem. There is a reason people pay, and will continue to pay, for premium services.

    • Hi Ben, thanks for commenting.

      I’m not calling anyone a bozo, quite the opposite.

      If, today, customers feel reassured by having specialist hosts, that is more than worth the $30 they are paying – many have been burnt in the past by completely awful hosts, while pretty much everyone is aware of the reputation WordPress has had for being insecure.

      I am saying that the value that the specialist hosts are perceived as providing may, in the coming years, be matched by offerings directly from the cloud giants. It makes sense for the cloud giant to invest that effort and pursue that market because your hosting requirements represent an important point-of-contact with you and your business.

      My point is that, actually, the specialist hosts do not do anything all that special but, because of those past problems, they get a disproportionate amount of credit for simply keeping WordPress up-and-running. I am suggesting that it will be possible for the cloud giants to achieve an even greater level of reliability and performance, at far greater scale and far lower cost.

      Not being bozos, this will noticed by many of the speciality host customers.

      • Good hosts are a you get what you pay for deal. I’ve had hosting at Host Gator, Blue Host, 1&1 to name a few of the names in known hosting. By the time we added up what everything cost for another I forget $15 or $20 we could go dedicated. We went with Codero, we have an i5 Server with two SSD drives as well as a HD, 20TB a month of bandwidth and the folks at Codero are not slouches. They KNOW hosting. I would bet dollars that one could take any WP host firm and Codero’s engineers will be more knowledgeable than they. Why?

        Two reasons. One, its not about the application its about maximizing the performance of the host platform and these folks are on par with Rackspace. Two, its about their infrastructure and it just FLIES. Its unbelievably fast. We slapped the sites built, Joomla’s and some eCommerce (including two Magento sites) on it and well… just night and day.

  16. I don’t think webhosts will die. I refuse to pay WPEngine’s prices when I can do things myself.

    I have two main hosts. I say this for sites that are my own. Clients’s sites that demand a specific webhost don’t count here…

    I have two main hosts, xxx and yyy. xxx is just for one site, the other 7 are on yyy.

    I prepay for two years, I get $4.99 a month. If I paid in a year advance only the $5.99 and monthly is $6.99.

    I like to manage my own sites, the files, updates and everything. As long as my server is up, I am happy. I have never had to wait more than 6 hours, I usually get a reply within the hour (most times within 30 minutes). The 6 hours was during a holiday weekend.

    I can’t justify in my mind WPEngine’s $29 price per month (cheapest plan).

    I am mentioning WPEngine because Ben Sibley’s reply (last one as I type this) has WPEngine in it.

    There will always be lazy admins who don’t want to do a lot. I like to actually administrate my sites.

    There will always be need for cheaper webhosts.

    Why is it “cheap” a bad thing? So many of those web hosts are really resellers.

    yyy webhost’s servers are good and stable, never had a big problem.

    All those WP focused hosts, what if I want to switch to Drupal? Joomla? or even have PHPBB? What then?

    • I get it, there are devs who want to control and manage their environment because they are good at it and enjoy it (and get paid for it). That makes sense.

      However, not managing your own hosting environment does not make a webmaster lazy. That’s like calling someone lazy for going to a restaurant instead of cooking their own dinner, or outsourcing any labor at that. I don’t care if I’m a chef, I don’t want to cook all my meals :)

      • Point is, that soon you will be able (some possibilities already exist) to manage your websites directly in cloud (Google, Amazon, Digital Ocean …) how you manage it today in wpengine, but these cloud providers offer better quality, performance and stability than most of hosting companies for lower price without restrictions what managed hostings usually have.

        If you have some time, you can already try to set up server on Digital Ocean via ServerPilot and you will easily recognize performance boost of your website for half or even lower price that you currently pay to your hosting.

    • You might want rephrase that. MOST of the operations that state they are dedicated to a software platform ARE resellers OR they have a co-located server where often it is managed by the firm where it is located. That is WHY the cost is high.

      As one poster noted, some might not wish to manage a server (its really not hard these days) and yes, you can slap whatever you please on it. But its also a good idea to LEARN as learning will allow one to better suit their clients and one will make better money at it.

      When things go wrong at a site, especially when a platform is dedicated to a given software package and support may cost $$$ how does anyone know that the “incident” really wasnt one at all.

      Generally speaking hosting is no different than anything else. You get what you pay for or less.

      Instead of “Twinkin” around to save a few peanuts go with a GOOD dedicated hosting firm and NEVER worry about it again. Its that simple. Codero treats clients like gold, almost like family. You can also use that to obtain clients. Let them know, “We dont use just ANYONE to host sites, we use dedicated platforms with real industry leaders and thats why the hosting cost is a bit more”

  17. “Thats funny. Obviously you have never worked in any enterprise environments.”

    Actually, I’ve worked in more than I care to have worked in..

    Shoddy Database 101 literally as is taught in first year college is breaking rules of normalization in a RDBMS. WordPress context? Exactly how much of the core source code have you studied?

    Uhhhh… what’s your point? Who cares what is taught in college? I’m not standing up for the way WordPress has it’s database setup. They’ve limited themselves in their belief that backwards compatibility is more important than cutting some loses for something that’s ultimately better. Just because core has thoroughly vetted queries doesn’t mean every theme and plugin does.

    Considering I submit patches often, I’d say I know WordPress pretty well.

    WordPress was already ported to .NET once and the results were 300-400% performance curve increase “out of the box”.

    Yeah, I’ve looked at It’s really not that impressive.

    Dont say never.

    WordPress core will never go to another language. They are too committed to backwards compatibility.

    There was talk here about it in fact at a “Word Camp” in Buffalo not to long back after the yip yap talk. I dont know if anything ever happened. They were talking about creating a fork of WP to either .NET or Java and reselling it. Some company from Korea or China. I never followed up on it.

    Exactly, it’s just a bunch of talk. Everyone says they’re going to make it better, or do something different, or blah blah blah. You didn’t even care enough to follow up on it.. so…

    • I am certainly newer to WP than say Joomla (which is a cow). WDotNet was not done to compete with WP. It was done to show that it can be done and that the performance of C# / ASP.NET is on a magnitude of 300-400% increase with the worlds most popular Open Source Application. If it were their goal to compete the project would be active .vs. 3.0 it was based upon.

      To me…. well, my personal thoughts are that all of these PHP CMS platforms are behind the curve already and will continue to wane as more and more enterprise level entities jump into the fray as they see gold in the horizon. Again, why wouldnt they?

      Amazon has a very different plan, I cant speak about it. But suffice to say many years back that did not hire part of Microsofts Core Visual Studio group (now Amazon employees) to convert the commerce venue to .net. They are in fact probably the entity that will be most formidable towards all open source projects that are latched into the PHP universe.

      Converting Joomla or WP to .NET is not a major feat as was shown. Its easily done. C# is a far more robust language than is PHP. But why would a third party do that in terms of commercial interests? No doubt that sort of thing would eventually turn to conflict with WP rights. If tons of the user base starts switching, not many options then.

      What is really needed in “Open source” is a new system that really leverages todays capabilities in web applications. Have you ever looked at Liferay? Check it out.

      Being committed to backwards compatibility is noble. But in reality how many of the plugins or themes have really done so? Regardless of those stats at least IMHO if backward compatibility end up being a roadblock to rapid advancement then its a boon not blessing. Microsoft essentially has proven this though Windows itself. Consumers will pay all over again.

      If the core doesnt see this and act upon it then they really are not very committed. It means they then know and accept its days are numbered and are just interested in riding the wave.

      They may have more insider info than I get. I know two coders (one I grew up with, one his pal) at Google, I know 3 developers at Adobe (and one that just left 6 months back now at Match), Microsoft I know several and many of those they let go have not exactly been silent. Amazon is a different gig as they are VERY serious about leaks of information but my brother worked tightly with them for quite sometime. Even locally since I made a regiment for several years of attending various user / coder groups and seminars get surprised at who one meets. I went to the places as they were excellent for getting jobs. Heck, we had for 5 years straight politicians paying for Joomla sites. Several LGBT sites, we’d done etc. all off referral from users groups.

      I’ve asked this question many times of folks. Larry Ellison. WHY did they buy up mySQL?

      Lots of people were FREAKING OUT if you remember when that went down, “Oh they are going to charge for it on and on”, “Oh mySQL they dont want compete with Oracle rdbms” etc etc.

      If one thinks like a giant such as Oracle w/ buckets of governance contracts and being one of few players left in the enterprise rdbms shakeout… Why would they spend what was it 1 billion dollars for mYSQL? Granted they snarfed up SUN and Java as well all totalling something like 8 billion?

      None of the above.

      Here is what I have heard, Its no secret that Oracle has never really fancied open source.

      Oracle wont ever charge for it, thwy wont re-license it, they wont sell it (they have had offers), as unification and the regulatory measures begin to unfurl (which we will see starting in the next administration here in the states as the platform is getting set for just that) they will pull the plug on it and they will tie up MariaDB in litigation.

      If you worked in enterprise IT then you know that business enemies and business friends reside on the same real estate. It depends which building you go to.

  18. Actually he is totally wrong. The opposite is happening. Cloud computing is what will disappear as we know it. Cloud computing is an utter failure and is not suited to application performance. The fog will wipe out public cloud computing eventually as demsification occurs and as applications improve. You will see companies all move to their own private cloud infrastructure. Amazon already knows this business is starting to peak and fail and they are already looking into a new drone business. Every business Amazon does lasts a few years and ultimately fails
    They ride on investment money until it fails and then Jeff Bozos simply comes up with a new business scam to get more investment money to keep the game going. Everyone knows this. Once mobile increases and densification increases you will see local m2m communications and the large cloud players will get thinned out by smaller ubiquitous cloud players. In fact you will likely even see all the web hosts merge their clouds into even more dense larger foundations than the big 5 and kill them off even sooner. The barbarians are already at the gate and Bezos already has his escape plan in place.

    • Who is this anonymous fellow “kingofnyct” who, six months after my article was published, turns up to post a rant in the WPTavern discussion and THREE increasingly deranged rants in the WPMayor discussion?

      I mean, seriously, what extremes of loneliness, procrastination and mania would have to possess a man to cause him to waste hours of his life, his precious irreplaceable life, tapping the same misinformed drivel into multiple websites?

      There are many angles that one could take to question a prediction looking ten years ahead, and many here made good points (six months ago), but it is hilarious that this moron doesn’t realise that most of the major WordPress hosts are actually built on the cloud he considers so unsuitable :D

      Also hilarious that he thinks Amazon is about to close down AWS, their most profitable and fastest-growing division with over $1 billion in annual profit.

      All of that pales into insignificance, however, when set against the personal tragedy of this poor, bewildered man, driven to pontificate but unable to squeeze out rational thoughts without spasmodically adding completely irrelevant buzzwords such as IOT and big data.

      Initially funny but, ultimately, a bleak warning of the perils of not having friends to hang out with on a Friday night.

      • Bye-bye Webhosting? Even COBOL lives on.

        Ideally, you’d be fielding replies to good posts, forever. Some rejoinders being more worthy than others, granted.

        I’d have to agree with Kingofnyct, that the IT biz model (Amazon, Uber, ad infinitum) does look suspect, Ponzi-like. Cloud, otoh, is potent tech; good for some things, not so much for others.

        Webhosting is like full-service gasoline. At first users were afraid of it, so we charged to do it for them. Then they decided they could do it themselves. Then they figured out they could drive off without paying. Now we see full service again.

        Chinese writing and their spoken language modified & guided each other, as they evolved. The same is true of computer & Internet hardware, software … and webhosting.

        We could deprecate hosting, live without it, but the world we have developed with it in mind & use, and it will (therefore) be the obvious solution to problems & needs we haven’t yet discovered.

        … but then, I always self-hosted. ;)


Subscribe Via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

%d bloggers like this: