Hosting the Leading Websites:
Making It on This List
The leading websites here are high-tech websites, those owned and managed at the top by institutions and people who are especially knowledgeable about information technology, such as standards organizations, industry members, specialized media, programmers, software and hardware firms, retailers, service providers, computer book authors, individuals, advisory forums, security firms, IT advocacy groups, and educational organizations. They are not likely to merely pass off the website stuff to a random contractor and, if they do, they look over the other person’s shoulders a lot, because, when it comes to computers, these IT leaders know what they’re talking about. Later, I may make similar lists for other kinds of leading websites, but first I wanted to know who people who live and breathe computer technology have hosting their own websites.
I’m not promoting my own host in this list (I wrote about it separately) and I’m not running a host on this list. I don’t know much on how to host. My Web host probably should be high in the list but is low, meaning less popular among the customers I researched, likely because it seeks a different customer base, customers whose content might be especially controversial. That’s an important mission, but this list tries to answer a different question. And I’ve never hosted any website. Even when I develop websites and view them locally, they’re not aimed at the Internet even temporarily. As far as I can tell, my host, whom I pay at its regular prices, displays them very well.
If you know a fabulous host, the way to add it is to persuade high-tech customers to use their (or your) website hosting and then wait for me to discover this. I’m more likely to search for customers in additional market segments or similar to those I’ve already researched, rather than re-research the same particular customers over and again. I also think I’ll let at least a few months pass between batches of customers, so that there would be time for trends to change. After researching over 350 websites one by one in the initial research before I posted any results, I notice that the list of Web hosts is beginning to stabilize. This probably does not make it easier for a new Web host to make it up here, but I think it maintains the integrity of my published results.
I tried to get all of the very best Web hosts, and did it in November–December, 2016, but one that had only a few customers surprised me, because I’m sure it’s very good at hosting and is well-known. For example, it was once hired by a foreign military because it could defend itself against cyberwarfare better than that nation’s military could. I don’t know why it had few customers among those I researched. But it did, so it’s low in the list. My guess is that it’s just not a computer firm customers think of when it comes to hosting.
Maybe I’m all wrong and these hosts are the worst. I wouldn’t rely on all of them for my needs, even though I compiled the list. Maybe these high-tech people were bored and asked their local pizza shops to find any host they could. But I doubt it. Airline pilots don’t just hang any old outboard motor on their planes and then pick up passengers. Not usually, anyway, and I think that’s true of people who can do their tax forms in hexadecimal quartets. They tend to be picky about the things they know about. When you want to climb Everest, it helps to ask someone who’s done it. That’s basically what I did here.
I didn’t list the customers’ websites I researched, because if someone wants to copy my research data and brand it as their own, I would rather that they did their own research from scratch. They might develop a better method. If so, they would earn my congratulations.
If a larger entity owns or controls a host’s domain, often the controlling entity — that’s the parent — decides who supplies services. I suspect that commonly includes hosting. If the parent is mainly expert on computers, then I consider mainly the parent and not so much the smaller unit. If the parent is mainly expert on something else instead, I omit them both. An example of the latter, that I would omit, would be an information technology department of a government or of a university with mostly non-computer majors. The main limitation on this is whether I find out that an entity is part of a larger entity, since I did only limited research on organizational parentage.
Game sites were excluded from my research. If a game producer mainly develops the entertainment and leaves the programming to someone else, they’d be irrelevant here. If they’re the other way around or they do both on major scales, they would belong, but I don’t know how to tell which firm works which way. I’d have to be inside one of these companies to know how they organize their work and I’m not. And I wouldn’t be surprised to find that many that are interactive multiplayer games are self-hosted, because I expect they’d want hands-on control of their servers. At any rate, the firms that don’t mainly do games reveal a good selection of Web hosts.
I left out hosting services as customers. I include a wide range of customers with high technical skills, except generally for Web hosts as customers, because I assume Web hosts host themselves. That wouldn’t add useful information to the results here, so I focused on customers likely to choose Web hosts and share them. That’s useful for us.
Bulletproof hosts are absent, happily. Sites that offer tools for illegal cracking of computers, copyright infringements, child pornography, and other content targeted by law enforcement agencies are not of interest here.
Not all kinds of customers are helpful for this research. I think many customers would not think deeply about a choice of hosts. Instead, they’d go to whomever can make them a website. I am thinking of adding a couple of customer market segments whose host choices might be helpful to us, and maybe I’ll do that research and include the results someday.
Many designers recommend their own Web hosts. I think designers often choose hosts based on financial deals, such as for hosting many websites in bulk, and the availability of customer support, but artists who know graphics would probably not be as knowledgeable on technical issues.
A national bias is present in the choice of Web hosts. I read English. I don’t know other natural languages and don’t want to rely on translation services. I generally avoided customers who primarily serve only non-English parts of the world. (I do include customers who are bilingual or more multilingual as long as that includes English. I include hosts regardless of language.) The United States probably leads in information technology and that will be reflected in the research results here. And registrars for some country code top-level domains (ccTLDs) do not help identify Web hosts. However, I don’t refuse to consider domains with non-English ccTLDs or websites written without much English if I otherwise know what it’s about. And, once a host is identified, I do not apply a national filter to it or to its rank. For a host’s URL, I give an English-language portal, if found.
Advertising did not corrupt my choices, and I don’t intend to let it. I want income from advertising, but if this website is good for advertisers it will be because interested people visit this website. Their visits are likely only if my content is trustworthy. I do not raise or lower a website host’s position in this list according to advertising or other payment or in pursuit of either.
Even a host I don’t like could be in this list. If it’s there, it earned its way by meeting my standards for this list. I kept my opinion out of that matter. The criteria for inclusion are the same.
High-tech sites usually could host themselves, or so we might expect, since they’d know how to select a staff and software. But, likely, it would be too expensive for the volume of traffic they’ll get, so they share hosting with other customers. Some of these customers do self-host, but in general I avoid researching customers who I anticipate will do that. If I don’t anticipate it and it turns out they self-host, I don’t count the self-hosting in ranking web hosts. I count their other customers. All of the Web hosts on the list are serving customers besides themselves.
Very high levels of security are demanded by some customers. They might require the ability to withstand the scale of denial-of-service attacks that a national government might launch against an international sovereign enemy in cyberwarfare. This list probably doesn’t adequately address that. I suspect researching those customers would be pointless, as they probably have their website hosting provided by themselves or by government agencies, and those services probably are not available to nongovernmental customers.
If I didn’t find out the host, then I left it off the list. There were very few like that and, in total, they had very few customers from my selection.
To define host: I mean the ‘host that serves websites’, not the ‘host that is the domain with potentially a website and subdomains’. Example.com can have a website and therefore just the domain example.com is itself called a host, but in this website I’m generally referring to the organizations and people who provide hosting for a website at example.com and usually other websites.
Stability likely marks a customer’s choice of Web host. I wish the research had taken only a day, but no, it did not and it’s okay that research was done in stages. For any single Web host, the research was likely spread over several weeks (in overlapping time frames). The chance that a customer changed its hosting in that time is negligibly small. Considering the substantial size of the customer base already researched, the effect of a Web host change in a short term on the whole list is vanishingly small.
These customers likely evaluate their hosts more or less daily, and probably more closely than less-skilled customers do. I didn’t evaluate the hosts directly, for the most part, but, considering the selection of customers, this list is evidence of the quality of these website hosts.
Your Bottom-Line Choice
Doing your own hosting is not recommended unless you have enough funds and experience, you don’t mind an amateur presence, or you thrive on risk. You can get your feet wet with high-quality highly reputed free software, a home computer (in my experience $50–150 used or free if found in the trash) connected to the Internet, and a minor experimental website, but many websites, especially brand-name, controversial, and income-earning websites, get attacked and hijacked and you’re going to need good security against hostile parties who are creative and able to change their ways remarkably often, which forces you to keep up to date with them. You’ll need to be able to write programs and patches fairly often, possibly every day. The more complicated your website, the more work you’ll need to do when self-hosting. If you really want to self-host, one compromise is for you to host your own website and also use a separate Web host for the same website, although that may require special arrangements for the DNS records and maybe that you devote time to maintaining two identical websites.
Have a second host, at least have its name in your files and have already checked their terms and offers, in case the first one is unsatisfactory and you want to make a rapid change. Keep an offline up-to-date copy of your website and its underlying software requirements in case you can’t recover your files from your first Web host. However, most of the customers I researched use only one website hosting firm at a time per customer, apparently. Some major sites use two or maybe more Web hosts simultaneously, in case one fails temporarily or so that distant visitors can have a host near them for faster download speeds, although having multiple hosts increases the website’s maintenance burden. For my own websites, I selected a backup Web host perhaps a decade ago, but my current Web host has been reliable enough that I don’t think I ever contacted that spare host. I do keep current offline backups of my websites. If you do switch hosts, it can take time. For *.com domains, it can take up to two days to start receiving traffic at the new host, because that’s how long it takes for your new IP addresses to propagate throughout the *.com DNS servers worldwide. For other top-level domains, expect a similar kind of propagation delay. As far as I know, the only thing you can do about it is to have simultaneous hosting before you have to turn one host’s service off. To sample propagation data, visit whatsmydns.net.