OK, so, if you look at what has set google apart you see three different things. Hardcore CS savvy is one of them, which led to pagerank. Then you have solid software engineering talent which led to the implementation of pagerank through map reduce and sharding. And combined with that you have revolutionarily different data centers. To google a data center is a single unit at many levels (application level, servicing level, manufacturing).
Google manufactures what are arguably single data center sized computers, and makes the software to run on it. This sort of vertical integration is a huge advantage to google, because it gives them a huge quality/price/performance benefit, allowing them to leverage CPU cycle more effectively, more efficiently, and more inexpensively than their competition. Can you come up with an example of a company that scales the same way that google does to the same degree?
Of course there are lots of other search engines that use pagerank. Sharding is of course nothing new. Map-reduce is a smart way to increase throughput of a cluster, but it's not unique (after all there is a popular open-source Hadoop project). I'll give you the data center advances though :)
Exactly, there are a lot of bits and pieces duplicating things that google did in the late 90s but there's nobody out there who's doing all those things at once. Which is surprising because there are huge business advantages to doing so. But it requires a level of execution that only google has been able to pull off so far.
Edit: on the one hand it's an impressive achievement on google's part and a testament to their exceptional competence. On the other hand it's a bit disappointing because it means that google is that much harder to compete with and diversity in the marketplace is always a good thing (no matter how benevolent the rule of the monopolist of the moment might be).
I think many would argue that their shipping container full of motherboards style data centers may have hampered them when it came to an IaaS offering. They didn't leverage VM technology at all did they? That certainly created a bias that led them to GAE.
Amazon, Facebook and Microsoft certainly must be operating datacenters on a massive scale and sophistication. They are probably more traditional thus less interesting to tech writers.
Eerie indeed given their existing vast bulk of personal data from emails to browsing habits. Now they want to be the portal to the entire web for people via Chromebooks and hold all their data. That makes the cynical part of me glad that Google+ hasn't been overly successful.
Search, Analytics, Plus, Mail, Android. Search & Analytics & Plus have allowed them to track you around the web. Mail & Plus have provided them with your real name, your relations/contacts and other personal details. Android gives them your phone number and location.
I believe part of the reason they are so steadfast in having real names for Plus is because Mail wasn't 100% successful in getting the actual names of users.
Everyone of their services is designed to gather a specific kind of information about the user. (Play a game and predict their next service based on the kind of information you'd expect them to want of you. I'm going with world wide payment systems.)
9 comments
[ 4.0 ms ] story [ 32.8 ms ] threadGoogle manufactures what are arguably single data center sized computers, and makes the software to run on it. This sort of vertical integration is a huge advantage to google, because it gives them a huge quality/price/performance benefit, allowing them to leverage CPU cycle more effectively, more efficiently, and more inexpensively than their competition. Can you come up with an example of a company that scales the same way that google does to the same degree?
Edit: on the one hand it's an impressive achievement on google's part and a testament to their exceptional competence. On the other hand it's a bit disappointing because it means that google is that much harder to compete with and diversity in the marketplace is always a good thing (no matter how benevolent the rule of the monopolist of the moment might be).
Amazon, Facebook and Microsoft certainly must be operating datacenters on a massive scale and sophistication. They are probably more traditional thus less interesting to tech writers.
I believe part of the reason they are so steadfast in having real names for Plus is because Mail wasn't 100% successful in getting the actual names of users.
Everyone of their services is designed to gather a specific kind of information about the user. (Play a game and predict their next service based on the kind of information you'd expect them to want of you. I'm going with world wide payment systems.)