What happens to startups once they are acquired by Google?


When companies get acquired by Google, initially many months are spent in the Operations area where the new startup is plugged into the Google infrastructure. The Google backend is very complicated, with many underneath systems. The most popular is BigTable and BigQuery that allows search for data. Other Google companies can search as well. These systems are not open for public because of scaling and threading requirements. The Google data is getting quite large. Already we are seeing that Google is only indexing the most recent data, ~ 5 years, and pushing all the Twitter content on top. This is because some very deep partnership in between Google and Twitter is in the works. Google is realizing that for long term strategy, it’s better to go with Twitter for the content because they have resolved their scaling issues. The systems are newer. It’s very difficult to upgrade legacy Google systems. Chances are in 100 years, we will only see very limited set of today’s data. However, there will be other companies that will continue to push relevant data forward as civilization needs change. Many large scale systems use map reducer and hadoop for their data clustering. However, this is a problem of big data. The data is impossible to retrieve so therefore summary tables in database design is a better choice. Find the relevant data, and abstract it in a summary table. There are periodic jobs that run and update the summary tables. The queries are quick and there is no reason to keep stale large data. After the summary tables are written, throw away the big data. Another issue with startups that are built on cloud and scaling large is the infrastructure costs. It’s not cheap to host a social network on a cloud infrastructure because of the amount of servers required, for example, the database servers, application servers, memory servers, and others. When Google acquired companies, they have to move these companies into their infrastructure as well. This adds to the operations cost.





Upload images with the blog.