I really like what Joyent is doing with illumos, dtrace, and other similar projects. That said...
They basically have let their "legacy" customers languish for a while, and after a pretty bad migration process a year or two ago, it was obvious that they really wanted to discard these old accounts - those boxes are running a version of opensolaris that's over 4 years old (snv_67).
Most of us paid quite a bit for our "lifetime" subscriptions, just to have it ripped up into a bunch of different parts that either get EOL'ed or sold off to another company (as was done with Strongspace).
Nope. Joyent is working with us to move the lifetime customers to some newer infrastructure, but they will be taken care of. Lifetime is a somewhat better story for storage, as you can consolidate 8U of power-hungry X4500 thumpers using 500GB disks down to a couple U of modern machines with 12x3TB arrays and a lot of RAM.
My account's server runs snv_121, so more like 3 years old, but still ancient. There was an attempt to upgrade us in December/January which was partially completed and finally "put on hold indefinitely". http://wiki.joyent.com/display/shared/Shared+Hosting+Mainten...
Mixed Griller here. I don't remember what the language was, but perhaps in this case "lifetime" meant the lifetime of the product, shared hosting. Personally, I'm happy with the accelerator, I never did like shared hosting and quit using it a long time ago except for a couple of sites that are gathering dust.
This is pretty common with consumer products. If something has a "lifetime warranty," that "lifetime" is typically a defined period of time (e.g.-seven years), not until you die.
For folks who are headscratching here, back in about 2007 (I remember it because I was getting into Rails) Joyent's precursor company TextDrive bootstrapped some hardware purchases by selling a (barely capable of running Rails, fairly rare for the time) shared hosting account. The kicker was that you'd pay about ~18 months upfront and it would be Lifetime. I don't trust my memory for numbers but somewhere between $150 and $300 if I recall correctly.
I actually used them as a hosting provider for a while, and followed the support forums. Supporting the Lifetime offering was a challenge pretty much from the getgo, because some folks were less than neighborly with their usage of system resources, partially out of ignorance, partially out of Rails playing very poorly with shared systems, and partially because you attract an interesting type of customer with this offering. The physical hardware had some faults and probably has not improved much over time.
Meanwhile, shared hosting for Rails is, well, not a very attractive option over the last couple of years, thanks to VPSes, Heroku, Amazon, etc etc. Joyent apparently wants to exit the business.
Of note: I remember somebody asking Slicehost to match the business model and Matt shot them down saying that it was too gimmicky for his taste and wouldn't be mutually beneficial. Slicehost eventually came up with a neat solution towards the same end: they were oversubscribed, so they sorted their waiting list by the amount of non-refundable deposit you were willing to make, giving them much-needed cash flow without committing them to service for forever.
Which (if you signed up at the end of 2006 and just got cancelled) comes out to about $7.50 a month. That's more than a EC2 micro instance on a 3 year reserved plan.
At the time, though, most people were just happy to find an easy solution to Rails hosting. I think TextDrive was the first provider in that price range that actually supported Rails. (Remember, this was before Passenger and all that made it easy to host Rails.)
Dividing a user's costs by the number of months of service they got isn't fair because it ignores the high risk they took on that the company would go out of business months after they dropped their $500 or $1000 or whatever.
Looking at my email I saw my Textdrive Lifetime hosting account was made in Sept 2005. I'm pretty sure I wasn't in on the first round either and I was second round "VC".
They also offered a lifetime package known as "Mixed Grill" and "3 Martini Lunch" that bundled the hosting, Strongspace (a file storage solution based on SFTP/rsync -- now spun off to a different company) and some collaboration suite (from Joyent).
I was (am?) also a TextDrive VCII customer. I signed up in March of 2005 and had a little home on a server called nelson. If my memory is correct it was around $450 or so and I hemmed and hawed over the decision. I was just getting out of the debt I had built up after the post-college/dotcom recession and it was a lot of money for me. However, I knew that I was going to continue programming for my career and so I saw it as an investment in them as well as myself. In 2008 I was asked to move to a Joyent shared accelerator (drake) and everything continued along.
I saw it as a genius business move for them: they got upfront capital in exchange for a promise to offer a commodity that their customers would naturally outgrow. Those customers would likely be highly influential and/or become power-users. Furthermore they would be able to grow bigger than they would otherwise without having to take VC money (which at the time was harder to come by and the VC had a lot more power/control). Now we have things like Kickstarter so the idea doesn't seem so strange anymore.
For abuse, which seems somewhat reasonable. And for the other people, they offered their money back, which is probably the best way to solve this kind of thing.
Matt, I was a very early slicehost customer, and just wanted to say thanks for the years of great service and straightforward business. I really loved what you did and hope you're continuing to do well. Cheers!
I paid $1399 for my TextDrive/Joyent "Three Martini Lunch" hosting, about six years ago. Which works out around $20/month. In retrospect, paying that amount up front doesn't seem "too good to be true", unfortunately.
There were actually a number of "VC" phases with TextDrive, starting with the "VC200" in 2004 that raised $40,000 from 200 pledges of $200 to acquire and set up the hardware--
That was followed by a number of upgrade offerings ("Mixed Grill", "3 Martini Lunch") in subsequent years, extending into the Joyent merger/acquisition, that required a larger lump-sum payment.
Their pitch on lifetime service? "How long is it good for? As long as we exist."
Now, you can argue in strictly financial terms that those initial customers (of whom I'm one) have probably got their money's worth, and that the hosting landscape has changed sufficiently that the lifetime products are peripheral to Joyent's main business. The counterargument is that Joyent not only acquired TextDrive's customers, but the goodwill surrounding Dean Allen's original venture, and has traded on that goodwill ever since. Clearly, they feel that's not worth much these days.
Furthermore, giving people who've had eight years of not having to think about hosting options just 80 days to migrate, with explicit notice that their servers will be shut down and wiped on October 31, strikes me as pretty cheap.
As far as getting one's money's worth, the "VC-like" pitching makes that odd. Obviously it wasn't an actual VC arrangement, but it had some similar aspects: you take a risk by giving $X now to buy a "lifetime account" with some startup that might be bankrupt next year. In return for helping to fund them and accepting that downside risk, your potential upside is that if they do succeed and exist long-term, you get what would in retrospect be below-market-priced service as a reward for your early support.
It doesn't make much sense to say that, hey, thanks for the "VC" investment when it was a risk to buy from us, but now that we're successful, you've gotten your money's worth.
I switched across to the Mixed Grill in May 2006 (right around when John Gruber quit Joyent to do Daring Fireball full-time, I remember). It cost $499USD.
Did I expect more than 6 years? Yeah, I took them at their word - "as long as we exist".
Their behavior here is embarrassing and labels them as not trustworthy in the future. Pacta sunt servanda. The other side ending up potentially getting the better end of the deal, a nowhere near certain thing here from a 2004-6 perspective where their customers took significant risk taking the deal, is not an equitable reason to terminate the services.
>We're pleased to announce a special offer that combines three great products in our family — TextDrive, Strongspace and Joyent — available for a one-time payment of just $499.
>What do you get?
>TextDrive's spectacularly feature-rich web, mail and data hosting; oceans of backup room at Strongspace; early access to the future of web-based organization, communication and productivity through the Joyent suite of interconnected applications. Perfect for a smart small business or smart individuals of any size.
True. You'd have sue and first get the arbitration clause revoked. My understanding is that courts are getting more amenable to this.
Dittos restrictions on class actions.
My policy is that I will generally reject arbitration / class action limitations and/or take my business elsewhere.
Of course, it's always possible that the arbitrator would find for the claimant, but such clauses usually (didn't check here) disallow class actions, and arbitrators are historically biased toward the entity requiring their services (e.g.: Joyent in this case).
Seriously. We are looking at hosting right now and are considering Rackspace and Joyent. In fact, I am one of the lifetime hosting people screwed over by this move, but this makes it an easy call. No way we are trusting Joyent with anything.
Rackspace are amazing. The only gripe I have (which isn't minor, I might add) is that you can only buy machines in tiers. If you want a machine with high network throughput (say, like a proxy), you need to purchase lots of cpu, ram and disk space that you have no need for.
Yes, I really like Rackspace. Great customer service! And the tier setup makes sense to me. It's Smith's Division of Labor principle at play. Sure, it's technically possible to adjust each feature separately, but you add so much complexity at every layer that it's a wash economically.
In the same vein, I can buy a Honda Civic for $10K, but a car with custom specs isn't $12K, it's $some_orders_of_magnitude_more, even if it's equivalent to a Civic in almost every way.
I fully agree with that. It's just something to keep in mind before building on their platform. It's not for me to decide how Rackspace should run their business, but for me it simply means I can't use them even though I want to. If only they kept a few "asymmetric" configurations as options, like Amazon does. But, alas.
I paid the extra $$ for lifetime (and did it back when I was a grad student and had basically zero money) because I didn't want to have to think about hosting any more. Now I have to think about hosting again. Ugh.
On a related note: where is good for hosting two small websites? I'd like to be able to log in with SSH and get an actual Unix shell (and not have to pay too much, as the websites are neither large nor popular).
If all you need is a shell plus LAMP, Dreamhost is pretty decent shared hosting. (Their Rails support apparently works, but Rails-using customers seem to have a tendency to suck up a lot of resources that other people could be using more moderately.)
I have a small one from them as a NS server and has no issue at all. However their bandwidth allocation (40G for 256m plan) is not enough for pretty much anything else.
As long as it's not terribly mission critical I'd suggest a VPS. Since you have a virtualized server you get everything you mentioned (and the responsibility of keeping the damn thin up to date). Take a look at http://lowendbox.com You're likely to find deals in the $20/yr range for small instances with ~128MB RAM. Expect 99% uptime and you'll be pleasantly surprised with the third nine.
If you chose a supplier at random, they yes, some of those will probably disappear after a few months. But if you spend some time digging around a little, you'll find things like "Best Provider of the Year" type posts (e.g. http://www.lowendbox.com/blog/best-low-end-providers-in-2011... ). These tend not to be fly by nighters.
Consider starting with that list, and look for recent offers by them. Also read the comments on each offer looking for irregularities. Lowendbox has a decent community, they know when an offer smells, or when a fly-by-nighter drops in. Granted it won't be perfect, but you can reduce the risk of a provider disappearing all of a sudden with a little bit of reading and digging.
Two in particular I find trustworthy: VooServers and Evorack
But beware, many low-end-box type providers use stock standard off-the-shelf tools and software, including WHM. So if you have a religious disagreement with passwords in cleartext emails, you should steer well clear of lowendbox/WHM combinations and go for a highend supplier instead.
Or perhaps a more practical method, don't use a secure password when you sign up, and make it more secure on the VM itself when you log in for the first time
I didn't end up playing with any low end VPSs because none of the providers I evaluated offer users the ability to clone VPSs, which is one of my requirements.
That's the nature of the business I'm afraid. But I have found some reliable hosts through the site and some very established hosts used to make offers before the siteowner left.
Regardless of where you host, if that's a problem for you, just set up a free account at Pingdom to hit it every 1 minute, and get uptime/response time monitoring as a bonus. Easy peasy.
Yes, I just shamelessly put in an affiliate link, but the service really is very good. It's shared hosting but you get SSH, a smart admin panel and very reasonable prices for your memory. After a few years using it, I'm also happy to report that their technical support is excellent - timely, helpful and knowledgeable.
What I disliked about webfaction is that the connection is proxied behind an nginx. But it is true that they have an excellent support. They answer almost inmediately. Also their Q&A was migrated to a stackexchange clone.
I prefer Linode, a little more expensive but you get decide even your own distro. (Yay for Archlinux)
Thanks for pointing out WebFaction. Of all the mentioned solutions WebFaction seems to me to be the closest thing to the old TextDrive shared hosting. I'm getting too old to want to get my hands dirty with managing my own VPS for a few simple PHP sites, some email accounts and my subversion repositories (and I can even switch to git).
Webfaction is also good if you want to run a no hassle Django installation. I initially got an account when I was learning Django and there weren't many Django friendly shared hosts out there. I have kept it to this day as their hosting panel is still pretty useful. They also have a bunch of automatic installs for rails, cherrypy, pylons, pyramid, subversion and trac etc. I've since moved on to do it myself on VPS but I still think Webfaction offers something different with their "framework" application approach.
NearlyFreeSpeech is very cheap and they even give you a shell with gcc (and I've run uploaded and run Go apps). They do require the applications to be static or CGI, though.
I really like working with Bluehost (http://www.bluehost.com/). They are amazingly cheap shared hosting plans targeted at non-administrators, but if you are willing to provide a copy of your government ID (as insurance against naughtiness) they will provide shell access. Their admins are very accommodating about letting me run my own stuff as long as it won't impact other users of the server.
Seconded. I've been using blue host for years with no troubles. Only downside is they keep lowering their rates, so I've got to call them every year or so to switch to the cheaper plan.
I host three sites on Linode, lowest plan at $20/mo. They're excellent, as just about everyone can tell you. It's a full install of a Linux distro of your choosing, any software you want. It's as good as having a Linux machine sitting in your office.
However, $20/mo might not be the lowest price you can find for two low-traffic sites.
edit - I should point out, I host three low traffic sites on one low-end Linode instance, so that' $20/mo flat for three sites. I also use it for compiling code (I can compile code anywhere from my phone!) and testing out new features for various bits of software, so it has more use than just hosting.
I use Linode for several low-traffic sites as well, and I'd recommend them.
However, there's one crucial aspect in which they're different than having a Linux machine sitting in your office (which I also have): for $100 I can put a 2TB drive in the local server. The equivalent space at Linode would cost $1000 a year.
If you have lots of data, Linode probably isn't a good fit. I'd look into Amazon S3 (which I suppose you could use with Linode hosting in front of it, but I'd probably just use AWS if I were relying heavily on S3).
Despite my post, I'm more into server administration than I am programming. Might be why I love Linode. I've helped a couple people set up their Linode sites, and tie it into S3 (and other services) for various uses. The thing I love most about Linode is that it's anything you want it to be.
For VPS, I recommend burst.net, specifically their XEN 1Gb ram plan. Their price structure is kinda strange, 0.5G is $7.95, 1G is $12.95, 1.5G is 24.95. So 1G plan is a no brainer. I have 2 of these, one in their PA Datacenter, the other in CA.
You might be interested in prgmr.com . VPS based on Xen. The owner is also a member of HN and wrote "The Book of Xen" a couple of years ago.
Also since you mentioned UNIX, maybe arpnetworks.com . As I recall they run Linux KVM VPS but offer OpenBSD and FreeBSD if preferred over a Linux guest.
Just to chime in with everyone else...I'm in the same boat. VC2/3 since Sept/2005. I started with them because they were the only host to offer Rails support. Now, my most pressing need is email support--transferring the main accounts, so that gmail can POP them off the accounts.
Edit: Obviously, any serious application would have been migrated off of Joyent long ago, but many of us have several small apps (Rails, Wordpress, etc) along with email accounts. I think it's more of an issue of migrating many small apps rather than a single large app.
I've used Slicehost (now Rackspace) successfully for a couple of years, but my pricing is probably terrible now. 256MB slice for $20/mth. I have a feeling I need to re-visit this...
I never asked much of it - all I really cared was that it was hosting that I didn't have to think about. Two WordPress sites, some backups of important files and the occasional scp around a stupid firewall or NAT box. And now they are canceling the lifetime service, or, put their ham-handed way: "your lifetime service will end on October 31, 2012."
Seriously. Someone in charge of PR actually wrote down and sent out an email with the sentence "your lifetime service will end on October 31, 2012."
Maybe they know something you don't know? Like perhaps they're planning on murdering you on October 30th. In that scenario, the sentence would make total sense.
Yup. Now as for damages, though, even though indeed they are supposed to cover your lost value and not the original price, the original price is going to set a point of reference that will influence how much you could reasonably convince any judge your lost value is. But say we compute it another way. Suppose you can buy equivalent service from Amazon for $100 a year forever. The net present value of that perpetuity at a 10% interest rate is $1000. You can easily see why if you imagine investing the $1000 at the 10% interest rate and using the interest each year to buy hosting: your principal would stay at $1000 forever. Now plug in numbers you actually think are feasible. And subtract attorney's fees. Good luck.
This actually seems fairly typical for most free or lifetime plans for things. Are there counterexamples of site/services that offered lifetime access that are still doing so 5 years later?
American Airlines started selling a lifetime pass in 1981, and is still honoring them today, although they're starting to crack down on certain abuses of the system.
Tivo has Lifetime subscriptions, where the lifetime is of the hardware but they transfer if you sell the hardware. They aren't always selling them as an option, but they always honor them, and have been for 10+ years.
I bought an original Tivo with lifetime service in 1999 and I still use the subscription today. 6 or 7 years ago they offered a promotion that allowed you to transfer your lifetime service to a new HD unit and that device is still running. The service has cost me an average of $1.28/mo.
In the end shared hosting with Joyent wasn't very good. It was slow, it was clunky and had an air of neglect. I have one site remaining on it I think. Most of my sites are on Linode.
What was worse was that Joyent changed directions, decided its then current customers weren't profitable enough (my guess shared hosting = higher per head tech support costs) and basically stagnated the service while it introduced new services. I think they did this a couple of times, and it has always made me relunctant to recommend them to friends (even ones looking for cloud services).
Sigh I hope StrongSpace will still honour the lifetime part.
BTW for readers who think lifetime account holders are being greedy, the point of the accounts was that when TextDrive/Joyent needed extra capital to expand, they offered lifetime accounts in return for quite a bit of cash up front. In part, they are where they are due to this clever bit of fund raising.
Joyent will keep supporting their Strongspace customers [which we, ExpanDrive, provide the service for - Joyent provides the hardware for their lifetimers]
I suspect this cuts right to the core of "why" Joyent would dump pretty low-overhead services and catch the flack over it. The hardware requirements to provide this shared hosting environment has plunged over the years; it's part of why I disagree with the "why did you EXPECT a lifetime deal" folks - because I always assumed that such a level of service would get cheaper and easier.
Except the one thing that doesn't get cheaper is the human power needed to administer the stuff and deal with the people. And as Joyent moves away from that sort of thing it becomes an unusual task and a drain on core business.
Which doesn't mean I think this is okay; the smart thing would have been to hand all of us lifetimers a lifetime bottom-tier slice. But I see why they'd want to be out of shared hosting.
The idea that the hosting would get cheaper over time is interesting. If you assume that the ongoing costs follow a Moore's Law type exponential curve (which may not be true, but seems like a decent rough approximation) then the total cost to provide a constant service for an infinite time is finite!
The real issue is, to paraphrase a football player, "What do lifetime mean?" Is it the lifetime of the product (typical in consumer goods), or was it the consumer's lifetime (i.e.-until they died)? If it was the consumer's lifetime, if they had a site where someone died and another person was going to take over, would we then be upset if Joyent started charging the successor?
Lifetime typically means expected life of the product (see: lifetime warranties). Death is really, really complicated, so I can't imagine anybody would actually tie their products to a consumer's death.
Yup, that's pretty cut and dry, and that's what I was looking for, so thanks. You guys that got hosed by this may have a breach of contract case if you'd like to take it that far.
If you sell a service, but decide not to provide it you should refund the money. Judging from the comments most of HN assumes if you "attempt" to provide the service you can keep the money.
Yes, I get it's a losing proposition for Joyent to live up to their obligations. So what? You either live up to them, or refund the money, it's really that simple.
Yeah, the comments seem to be filled with people saying that it's not practical to provide "lifetime", and people buying it should have known they didn't mean it anyway.
I look at this the same way I look at "unlimited" internet service. Unlike a lot of the tech crowd, I have no problem whatsoever with metered bandwidth. My only quibble is that if you advertise "unlimited" you had better actually provide unlimited. If you want to charge for overages or enact caps, you don't get to call it "unlimited".
Yeah, the comments seem to be filled with people saying that it's not practical to provide "lifetime", and people buying it should have known they didn't mean it anyway.
And that fascinates me. There's a clear information asymmetry here, and therefore a clear power differential. But so many people are instantly willing to blame the weaker party for being trusting, rather than blaming the stronger party for abusing their customers.
Of course, they're only doing that after it blows up. None of them are saying, "Stop using Hacker News! There's no revenue model, so it will all end in tears! Quit now!"
It's an interesting question. If I make a promise I know I can't fulfill, and you know I can't fulfill, is it still a promise? Is there a breach of trust when I end up not being able to fulfill it?
I'd say that yes, it's still a promise and a breach of trust even if everybody knows I don't really mean it. But I don't think there's a single obvious answer.
If you know you can't fulfill it then I believe it's all on you, in that you're lying.
Some fraction of people will believe you, and even those that don't can't know you can't fulfill it. There are plenty of legitimate life-of-the-company deals. One of them is buying stock. And there are plenty of companies that offer lifetime warranties. Only the person offering the deal has the information to know that they can't honor it.
And in this case, Joyent can honor it; it's not like keeping these servers running would put them out of business. They just don't want to honor it, so anybody who took this deal is so far correct in thinking that Joyent made a promise that they could fulfill.
Asking them to do it is fine, but you can't obligate them to do it. A contract can never force you to perform work, because that is slavery. It can force you to perform work or suffer some kind of penalty. In this case, I think the appropriate penalty is a full refund for all of the affected customers.
Are you a lawyer? I'm not, though I studied law for a while.
What you say may be true of common law. But I suspect the doctrines of equity may have other ideas. The return of funds is not the same as lifetime hosting. Those clients want specific performance and equity may yet grant it to them.
No, I'm not. And yes, the clients want specific performance, but do they have the right to get it? Seems like at most, the clients would be entitled to a refund plus whatever damages they could show (e.g. the cost of switching to a different service, plus any losses they take from other arrangements they made based on the promise of lifetime service). Could a court really order Joyent to keep providing this service against their will? Seems like restitution would be it.
Some people don't want damages or restitution. They want specific performance of the contract. In this case they may be in a position to get it.
We're not discussing labour laws here, the slavery argument doesn't apply. This looks to me like straight up contract + equity.
> Could a court really order Joyent to keep providing this service against their will?
That's what an order for specific performance is for.
I've emailed a self-described law nerd of my acquaintance. She geeks out on equity law, so this sort of case is right up her alley. I'll be interested to see what she thinks.
"your lifetime service will end on October 31, 2012"
If you're writing that sentence and don't see a problem with it, there's no helping you. Look, many of us have been there and sold "free" things to early customers who become a pain later on, but you have to honor it. Better yet, never sell something "lifetime" without at least some kind of low recurring fee to cover nominal costs.
Thank you: just the idea of someone sitting down and typing that phrase in amazes me. Some kind of token refund and a slightly longer switch-over period could have turned this into a PR positive.
As annoyed as I am with all of it, I'm more interested to know how to move off of their products and services now.
My lifetime account is where I have several important email servers, and I don't know how to migrate them. All the email says is to contact Joyent about getting migrated to one of their other product offerings. However, if I am only given 2.5 months to sort all this stuff out, I DO NOT WANT to use Joyent services. I'd love to use GMail for Business, Linode or what have you, but I don't know where to start.
I tend to just use an IMAP mail client (like thunderbird) connect the old server, and new server, then copy from one to the other... this doesn't take care of multiple users and their accounts/passwords though.
Set up an IMAP connection, copy everything over to your local machine, connect via IMAP to your new mail host and move everything over to it. I usually use Gmail for Business but be aware that this now has a 10 user limit on the free account.
When Connector shut down, I switched to Fastmail, which had a nice little "fill in the credentials and press the button" migration process. Been with 'em ever since.
I too switched to FastMail a year or so ago for my email, as I found Joyent email to be slow and didn't have any spam filtering functionality. They have an IMAP/POP email importer which will suck in all your email and folders from you previous provider. Worked like a charm.
I do not want to be a jerk here, but the only reason I joined was the promise of "lifetime" - they can NOT just cancel our accounts does it does not suit them. Please contact me webmaster @ opendomain ORG if you would like to join the class action lawsuit.
Wow! I have received over a dozen emails in the first hours that want to join in the lawsuit. We have enough to form an offical class and will be able to get Joyent to abide by the original terms. Please contact me soon if you want to participate!
Had an account with them back around then because their site made them seem technically proficient, but dealing with their customer service was like dealing with bad sys admins - grudging help and lazy. There were also a lot of unfinished nooks and crannies in their system. An overall lack of craftsmanship and care. The email brought back not so fond memories and is indicative of their attitude in all their interactions. Partly you are paying for company culture and theirs is not good.
I wonder why Joyent thinks they can get away with this. "Q: How long is it good for? A: As long as we exist." seems to be a pretty good definition of "lifetime." I would think they could only get out of their promise via bankruptcy. Maybe Joyent is trying to substitute for legal bankruptcy with moral bankruptcy.
"I wonder why Joyent thinks they can get away with this"
"As long as we exist"
Possession is 9/10ths of the law.
Nobody is going to file a private lawsuit to enforce this (assuming there is even a leg to stand on to enforce it I haven't read the exact contractual promise..) And the other option would be a class action lawsuit where there would have to be enough class members to interest an attorney.
(Not relevant in this case but if a company was acquired the new company may have not acquired all the previous companies legal obligations but simply setup a new entity and assumed some of the assets, liabilities and obligations but not everything. When I sold a company that is how it was handled.).
Although it still wouldn't be a cost-effective use of time, a sufficiently peeved customer could file a small-claims suit for a return of the fee (or some portion of it).
Actually not a bad strategy assuming the venue is convenient.
A corporation has to be defended by an attorney while an individual can be pro se.
As a result the corporation would typically want to settle rather than spend money to go to defend in court. My guess is as long as the person filing the suit didn't blog about it and bring attention to the action beforehand there would be incentive to settle this out of small claims court. The amount settled for could easily be the value over time of the "lifetime" benefit. If the company choose to go to small claims court and lost it would certainly not be to their benefit as it would show others what was possible and set some precedent.
Keep in mind that even if the company feels their contract is airtight they might decide to settle simply to not have to pay legal fees in many cases.
Why wouldn't somebody sue? It sounds like this service cost a minimum of hundreds of dollars, which is enough to be significant, but still small enough to fit within small claims. If I were one of these customers I'd certainly consider it.
Not necessarily. They could simply settle with those who sue and not let it get to court. I bet not even 10% would even sue in the first place, so they could still come out well ahead. I wouldn't be surprised if they accounted for that in their planning when doing this in the first place.
Had a similar experience with an old host, promised "lifetime price freeze". A couple of years later they started charging me more as they had 'moved me to a new server' which I fiercely contested. In the end we settled for a compromise.
It's always amusing when companies (or people) contradict themselves to rationalize their decisions. The ongoing costs of this service are too large for us to sustain! By the way, only a few obscure people are using it so it will impact virtually nobody!
The service may have high fixed costs. If they have one person maintaining something that's used by very few customers the cost per customer would be very high.
Much suckage. Lifetime means lifetime. It wouldn't be too much out of the way to provide a small host for "lifetime" users. I have complained to joyent.
I wonder why they are this dumb. It's like buying bad press. Did they really think this wouldn't get out and make them a mockery? Would it really cost that much to continue providing something to these people forever, so they wouldn't have to say the lifetime is over? Dumb dumb dumb.
Seems a little odd, surely with the cheaper RAM, far more powerful CPUs and inexpensive disks these days, they could spend a few $$$ and burn a couple of U , keeping their customers happy?
I don't know how many servers they would have needed in 2007 to provide the services they sold, but surely now they could consolidate all those users onto 1/4 the machines?
I think the real issue is that Joyent no longer provides any form of shared hosting and they don't want to. So they have no equivalent service to migrate the old lifetime customers to. They'd probably lose quite a bit of money if they migrate lifetime customers to their cloud.
273 comments
[ 5.0 ms ] story [ 182 ms ] threadThey basically have let their "legacy" customers languish for a while, and after a pretty bad migration process a year or two ago, it was obvious that they really wanted to discard these old accounts - those boxes are running a version of opensolaris that's over 4 years old (snv_67).
Most of us paid quite a bit for our "lifetime" subscriptions, just to have it ripped up into a bunch of different parts that either get EOL'ed or sold off to another company (as was done with Strongspace).
Any intention to EOL the "lifetime" plans on your end?
Q: How long is it good for? A: As long as we exist.
They should of given you at least 15 years free I think.
http://web.archive.org/web/20060203030930/http://www.textdri...
Q: How long is it good for? A: As long as we exist.
Seems pretty clear cut to me.
(Not sure this is accurate, though. I did find it repeated in a forum discussing Joyent in 2006.)
http://web.archive.org/web/20060202181857/http://textdrive.c...
I actually used them as a hosting provider for a while, and followed the support forums. Supporting the Lifetime offering was a challenge pretty much from the getgo, because some folks were less than neighborly with their usage of system resources, partially out of ignorance, partially out of Rails playing very poorly with shared systems, and partially because you attract an interesting type of customer with this offering. The physical hardware had some faults and probably has not improved much over time.
Meanwhile, shared hosting for Rails is, well, not a very attractive option over the last couple of years, thanks to VPSes, Heroku, Amazon, etc etc. Joyent apparently wants to exit the business.
Of note: I remember somebody asking Slicehost to match the business model and Matt shot them down saying that it was too gimmicky for his taste and wouldn't be mutually beneficial. Slicehost eventually came up with a neat solution towards the same end: they were oversubscribed, so they sorted their waiting list by the amount of non-refundable deposit you were willing to make, giving them much-needed cash flow without committing them to service for forever.
http://discuss.joyent.com/viewtopic.php?id=13148
They also offered a lifetime package known as "Mixed Grill" and "3 Martini Lunch" that bundled the hosting, Strongspace (a file storage solution based on SFTP/rsync -- now spun off to a different company) and some collaboration suite (from Joyent).
I saw it as a genius business move for them: they got upfront capital in exchange for a promise to offer a commodity that their customers would naturally outgrow. Those customers would likely be highly influential and/or become power-users. Furthermore they would be able to grow bigger than they would otherwise without having to take VC money (which at the time was harder to come by and the VC had a lot more power/control). Now we have things like Kickstarter so the idea doesn't seem so strange anymore.
http://photodude.com/2004/06/01/textdrive-or-how-to-raise-40...
That was followed by a number of upgrade offerings ("Mixed Grill", "3 Martini Lunch") in subsequent years, extending into the Joyent merger/acquisition, that required a larger lump-sum payment.
Their pitch on lifetime service? "How long is it good for? As long as we exist."
http://web.archive.org/web/20060202181857/http://textdrive.c...
Now, you can argue in strictly financial terms that those initial customers (of whom I'm one) have probably got their money's worth, and that the hosting landscape has changed sufficiently that the lifetime products are peripheral to Joyent's main business. The counterargument is that Joyent not only acquired TextDrive's customers, but the goodwill surrounding Dean Allen's original venture, and has traded on that goodwill ever since. Clearly, they feel that's not worth much these days.
Furthermore, giving people who've had eight years of not having to think about hosting options just 80 days to migrate, with explicit notice that their servers will be shut down and wiped on October 31, strikes me as pretty cheap.
It doesn't make much sense to say that, hey, thanks for the "VC" investment when it was a risk to buy from us, but now that we're successful, you've gotten your money's worth.
Did I expect more than 6 years? Yeah, I took them at their word - "as long as we exist".
>One-time payment of $499
>We're pleased to announce a special offer that combines three great products in our family — TextDrive, Strongspace and Joyent — available for a one-time payment of just $499.
>What do you get?
>TextDrive's spectacularly feature-rich web, mail and data hosting; oceans of backup room at Strongspace; early access to the future of web-based organization, communication and productivity through the Joyent suite of interconnected applications. Perfect for a smart small business or smart individuals of any size.
>How long is it good for?
>As long as we exist.
Dittos restrictions on class actions.
My policy is that I will generally reject arbitration / class action limitations and/or take my business elsewhere.
Of course, it's always possible that the arbitrator would find for the claimant, but such clauses usually (didn't check here) disallow class actions, and arbitrators are historically biased toward the entity requiring their services (e.g.: Joyent in this case).
In the same vein, I can buy a Honda Civic for $10K, but a car with custom specs isn't $12K, it's $some_orders_of_magnitude_more, even if it's equivalent to a Civic in almost every way.
Broken promises leave a bad taste in my mouth.
Consider starting with that list, and look for recent offers by them. Also read the comments on each offer looking for irregularities. Lowendbox has a decent community, they know when an offer smells, or when a fly-by-nighter drops in. Granted it won't be perfect, but you can reduce the risk of a provider disappearing all of a sudden with a little bit of reading and digging.
Two in particular I find trustworthy: VooServers and Evorack
But beware, many low-end-box type providers use stock standard off-the-shelf tools and software, including WHM. So if you have a religious disagreement with passwords in cleartext emails, you should steer well clear of lowendbox/WHM combinations and go for a highend supplier instead.
Or perhaps a more practical method, don't use a secure password when you sign up, and make it more secure on the VM itself when you log in for the first time
YMMV.
http://www.lowendbox.com/blog/top-providers-2011q1/
http://www.lowendbox.com/blog/top-low-end-vps-providers-2012...
Here's a list of popular low end VPS providers at the LowEndTalk Wiki that is based on the same data used for the above lists:
http://www.lowendtalk.com/wiki/top-providers
Here are the popular low end VPS providers I made note of when I researched this subject recently:
http://buyvm.net/ (limited availability; see http://doesbuyvmhavestock.com/)
https://hostigation.com/
http://www.ramhost.us/ (owns http://tinyvz.com/ and http://tinykvm.com/)
http://www.securedragon.net/
In addition to LowEndBox and LowEndTalk, the following two sites had some good info about low end VPSs:
http://www.vpsadvice.com/
http://freevps.us/
I didn't end up playing with any low end VPSs because none of the providers I evaluated offer users the ability to clone VPSs, which is one of my requirements.
http://www.webfaction.com?affiliate=hammertime
Yes, I just shamelessly put in an affiliate link, but the service really is very good. It's shared hosting but you get SSH, a smart admin panel and very reasonable prices for your memory. After a few years using it, I'm also happy to report that their technical support is excellent - timely, helpful and knowledgeable.
I prefer Linode, a little more expensive but you get decide even your own distro. (Yay for Archlinux)
However, $20/mo might not be the lowest price you can find for two low-traffic sites.
edit - I should point out, I host three low traffic sites on one low-end Linode instance, so that' $20/mo flat for three sites. I also use it for compiling code (I can compile code anywhere from my phone!) and testing out new features for various bits of software, so it has more use than just hosting.
However, there's one crucial aspect in which they're different than having a Linux machine sitting in your office (which I also have): for $100 I can put a 2TB drive in the local server. The equivalent space at Linode would cost $1000 a year.
http://www.burst.net/xenvps.shtml
Also since you mentioned UNIX, maybe arpnetworks.com . As I recall they run Linux KVM VPS but offer OpenBSD and FreeBSD if preferred over a Linux guest.
Edit: Obviously, any serious application would have been migrated off of Joyent long ago, but many of us have several small apps (Rails, Wordpress, etc) along with email accounts. I think it's more of an issue of migrating many small apps rather than a single large app.
I've used Slicehost (now Rackspace) successfully for a couple of years, but my pricing is probably terrible now. 256MB slice for $20/mth. I have a feeling I need to re-visit this...
I think we had one site on Textdrive that was live, and quite quickly moved to another host for obvious reasons.
I'm glad it worked fro some people, overall a poor experience for us.
Seriously. Someone in charge of PR actually wrote down and sent out an email with the sentence "your lifetime service will end on October 31, 2012."
(IANAL, TINLA).
http://articles.latimes.com/2012/may/05/business/la-fi-0506-...
They ended this service in 2011.http://en.wikipedia.org/wiki/TiVo#United_Kingdom
They no longer honour this deal in UK (which sucks, because I loved it).
What was worse was that Joyent changed directions, decided its then current customers weren't profitable enough (my guess shared hosting = higher per head tech support costs) and basically stagnated the service while it introduced new services. I think they did this a couple of times, and it has always made me relunctant to recommend them to friends (even ones looking for cloud services).
Sigh I hope StrongSpace will still honour the lifetime part.
BTW for readers who think lifetime account holders are being greedy, the point of the accounts was that when TextDrive/Joyent needed extra capital to expand, they offered lifetime accounts in return for quite a bit of cash up front. In part, they are where they are due to this clever bit of fund raising.
Except the one thing that doesn't get cheaper is the human power needed to administer the stuff and deal with the people. And as Joyent moves away from that sort of thing it becomes an unusual task and a drain on core business.
Which doesn't mean I think this is okay; the smart thing would have been to hand all of us lifetimers a lifetime bottom-tier slice. But I see why they'd want to be out of shared hosting.
http://en.wikipedia.org/wiki/Landauer%27s_principle
Lifetime typically means expected life of the product (see: lifetime warranties). Death is really, really complicated, so I can't imagine anybody would actually tie their products to a consumer's death.
It has this Q&A item: How long is it good for? As long as we exist.
The footer states: a Joyent company
Nothing complicated, Joyent still exists. shrug
If you sell a service, but decide not to provide it you should refund the money. Judging from the comments most of HN assumes if you "attempt" to provide the service you can keep the money.
Yes, I get it's a losing proposition for Joyent to live up to their obligations. So what? You either live up to them, or refund the money, it's really that simple.
I look at this the same way I look at "unlimited" internet service. Unlike a lot of the tech crowd, I have no problem whatsoever with metered bandwidth. My only quibble is that if you advertise "unlimited" you had better actually provide unlimited. If you want to charge for overages or enact caps, you don't get to call it "unlimited".
And that fascinates me. There's a clear information asymmetry here, and therefore a clear power differential. But so many people are instantly willing to blame the weaker party for being trusting, rather than blaming the stronger party for abusing their customers.
Of course, they're only doing that after it blows up. None of them are saying, "Stop using Hacker News! There's no revenue model, so it will all end in tears! Quit now!"
I'd say that yes, it's still a promise and a breach of trust even if everybody knows I don't really mean it. But I don't think there's a single obvious answer.
Some fraction of people will believe you, and even those that don't can't know you can't fulfill it. There are plenty of legitimate life-of-the-company deals. One of them is buying stock. And there are plenty of companies that offer lifetime warranties. Only the person offering the deal has the information to know that they can't honor it.
And in this case, Joyent can honor it; it's not like keeping these servers running would put them out of business. They just don't want to honor it, so anybody who took this deal is so far correct in thinking that Joyent made a promise that they could fulfill.
What you say may be true of common law. But I suspect the doctrines of equity may have other ideas. The return of funds is not the same as lifetime hosting. Those clients want specific performance and equity may yet grant it to them.
We're not discussing labour laws here, the slavery argument doesn't apply. This looks to me like straight up contract + equity.
> Could a court really order Joyent to keep providing this service against their will?
That's what an order for specific performance is for.
I've emailed a self-described law nerd of my acquaintance. She geeks out on equity law, so this sort of case is right up her alley. I'll be interested to see what she thinks.
Of course ... IANAL, TINLA.
Keep an eye on http://skepticlawyer.com.au.
I once had a lifetime free checking account, I still have the advertising flyers from when I got it.
Well they took that away from me last year too.
I guess they just count on you getting a lawyer and lawsuit being more of a hassle than you just walking away.
If you're writing that sentence and don't see a problem with it, there's no helping you. Look, many of us have been there and sold "free" things to early customers who become a pain later on, but you have to honor it. Better yet, never sell something "lifetime" without at least some kind of low recurring fee to cover nominal costs.
My lifetime account is where I have several important email servers, and I don't know how to migrate them. All the email says is to contact Joyent about getting migrated to one of their other product offerings. However, if I am only given 2.5 months to sort all this stuff out, I DO NOT WANT to use Joyent services. I'd love to use GMail for Business, Linode or what have you, but I don't know where to start.
"As long as we exist"
Possession is 9/10ths of the law.
Nobody is going to file a private lawsuit to enforce this (assuming there is even a leg to stand on to enforce it I haven't read the exact contractual promise..) And the other option would be a class action lawsuit where there would have to be enough class members to interest an attorney.
(Not relevant in this case but if a company was acquired the new company may have not acquired all the previous companies legal obligations but simply setup a new entity and assumed some of the assets, liabilities and obligations but not everything. When I sold a company that is how it was handled.).
A corporation has to be defended by an attorney while an individual can be pro se.
As a result the corporation would typically want to settle rather than spend money to go to defend in court. My guess is as long as the person filing the suit didn't blog about it and bring attention to the action beforehand there would be incentive to settle this out of small claims court. The amount settled for could easily be the value over time of the "lifetime" benefit. If the company choose to go to small claims court and lost it would certainly not be to their benefit as it would show others what was possible and set some precedent.
Keep in mind that even if the company feels their contract is airtight they might decide to settle simply to not have to pay legal fees in many cases.
One of the few? Cool. It's not costing you that much to keep them on their lifetime plan. Surely its not worth the negative PR?
Not sure what Joyent is thinking.
I don't know how many servers they would have needed in 2007 to provide the services they sold, but surely now they could consolidate all those users onto 1/4 the machines?
What is this negative press costing them?