Friendster and MySpace failed because of scaling issues - but it was not scaling per se but bad execution: They choose the wrong technology, brought in the wrong CTOs who bought very expensive hardware (wrong, few expensive high end servers with expensive storage, instead several more low cost normal servers) and software with very high license costs (wrong, ColdFusion on dotNet (experimental) on Windows (high lic costs)). That's why both lost the race, they stalled development because of rewrites for a year (too long).
Edit: you can read about both project histories, and learn about them. Of course on gets down-voted for mentioning it.
As a counterpoint stack overflow is hosted on windows and did just fine. If your service is printing money you can easily fix tech problems (e.g. Facebook effectively rewriting php), if it's not you may flail around trying rewrites etc but the real problem lies elsewhere.
Tech people are quick to find technical reasons for failure but the reasons are usually elsewhere.
Stack Overflow seems very relevant to this conversation given how unremarkable their architecture is – their level of traffic is comfortably running on 4 MS SQL Server boxes:
>>> how unremarkable their architecture is – their level of traffic is comfortably running on 4 MS SQL Server boxes:
Wrong, and wrong.
First, their architecture is extremely remarkable. They are doing and mastering vertical scaling, down to every little details. Terabytes of RAM, C# instead of Ruby/Python, FusionIO drives, MS SQL instead of MySQL/PostGre, etc...
Second, that's at least 6 database servers, each one being more expensive than 10 usual commodity servers:
First cluster: Dell R720xd servers, 384GB of RAM, 4TB of PCIe SSD space, and 2x 12 cores. It hosts the Stack Overflow, Sites (bad name, I’ll explain later), PRIZM, and Mobile databases.
Second cluster: Dell R730xd servers, 768GB of RAM, 6TB of PCIe SSD space, 2x 8 cores. This cluster runs everything else. That list includes Careers, Open ID, Chat, our Exception log, and every other Q&A site (e.g. Super User, Server Fault, etc.).
I think you misunderstand my use of unremarkable. It's not wrong or bad but key parts are the same stack you'd have picked 1-2 decades ago – well implemented, values scaled up, yes, but something you could have shown people in 2000 and not had to explain more than that the hardware costs so much less.
I think that's good: SQL databases are very mature and you don't want to be exciting for your core business data if you don't get some major benefit to defray the cost. Boring is a delightful characteristic for data storage.
My bad then. I think that using a proven technology perfectly tuned and the way it's meant to be done, is remarkable. This should be promoted more, instead of the new and shiny.
I'm not sure you could get 1TB of memory and multi TB SSD drives anywhere in the 2000's, even for a million dollar. That makes a major difference in the ability to scale up. Data didn't grow, storing 1M user account always took the same space.
Agreed that data sizes would have been far more of a challenge – far more people used clustered services just because a single box only supported so much RAM, RAID arrays for IOPs and size as well as redundancy, etc. – so it's arguably become much easier for a growing chunk of the industry.
>believe they'd need a clustered NoSQL database or it wouldn't scale.
While for many developers, NoSQL might be overkill - Stackoverflow is a bad example. If you were any fast growing startup in the cloud, and you wanted to go the SO route it would have meant going CoLo. SO has set of machines with a nearly a terabyte of RAM - GCE doesn't even offer cloud machines with the same specs.
And even then their setup is far more fiscally expensive than something you could get done with 20 cloud nodes on some NoSQL solution.
One: run the numbers on the dozens of cloud nodes you mentioned. Is that really cheaper than renting some servers in a colo?
Second: how much time would they have spent rolling all of the data integrity, reporting, etc. features they'd have needed to add. I'm inclined to take them at their word when they say this was safer and cheaper given their resources.
>I'm inclined to take them at their word when they say this was safer and cheaper given their resources.
I never meant to claim that cloud was safer and cheaper. What I meant was, for the majority of operations, staying in the cloud with some distributed setup is likely more feasible that moving to CoLo (see GitLab).
1) Friendster had scaling issues. They insisted on keeping the x-degree-of-friendship-calculation which is quite resource expensive and doesn't scale. Instead making the software scalable, they brought in expensive exec who decided to through more and more very high end servers and enterprise storage on it. It eat their startup money, and they were constantly firefighting with very high page load time like 12+ sec, instead of advancing the site for at least one year. MySpace took over Friendster because of this.
2) MySpace culture and management was in trouble due Murdoch's News Corporation bought MySpace's parent company. Instead of investing News Corp choose the wrong path. Instead of improving the site, they decided to do a (in the end) very costly deal with MS to switch their ColdFusion stack from Java stack to dotNet stack incl very expensive MSSQL licenses and various other license costs. It turned out ColdFusion on dotNet was still in very experimental phase and MySpace suffered from the caused troubles a lot and bleed money like never before (which made News Corp very unhappy). MySpace website got little updates for at least one year while MySpace devs were busy with firefighting and switching backend. Facebook took over.
There are books worth reading.
3) Stackoverflow (like 2 years ago) run on less than a dozen of servers. It's several magnitudes smaller than social network services like (former) MySpace, Facebook, Twitter, etc. License costs often don't scale, you bleed through your startup money for little competitive advanced features or return. That's why successful startups often choose open source stack, look at Amazon, Google, (former) Yahoo, Facebook, Twitter, etc. - Perl, Python, PHP, Java, Ruby, MySQL, Postgres, Hadoop, etc. Stackoverflow shows it can be done with off the shelf COTS as well, if you keep the server license count low by wise decisions and performance tuning. But it's not like I could name dozends of successful startups that have a software stack like Stackoverflow. And even Hotmail, Linkedin, Bing run or used to run for the majority of their service-live on mainly open source software stack.
Friendster was mostly focused on Asia plus lost a ton of users around the same time Facebook gained a ton of users. I don't think it was scaling architecture that did them in. It was the market choosing the competition for the user experience plus what their friends were on. That's for most of the world. I have no idea what contributed to their failure in Asia since I don't study that market when it comes to social media.
EDIT to add: I recall the founder did say they had serious technology problems for a few years that affected them. I'm just thinking Facebook spreading through all the colleges & moving faster on features was their main advantage.
Asia? You came a few years to late... Friendster was US focused, at least until it lost. Then the probably pivoted, but I wrote about the first phase of Friendster (the one everyone in US remembers who was active online in early 2000s).
Hmm. It peaked in Asia with most users in Asia operating from Asia. You don't think the Asian focus could affect its marketing strategy for Americans? Facebook's college angle worked really well after MySpace's express yourself in cluttered pages worked before that. Im not sure Friendster understood our userbase enough to keep it.
Edit: you can read about both project histories, and learn about them. Of course on gets down-voted for mentioning it.