Google F1, The Fault-Tolerant Distributed RDBMS is targated to be replacement for MySQL. This claimed to be perform better like serving Ads from crap servers. Those who are not sure why I started with an attacking mood, reading this post will make easy to digest this article.
The Official Story behind Google F1
Google Research has developed a new F1 database which said to be fail-safe and scalable NoSQL database, but it offers the advantage of a SQL database. In F1 there is a relational database with a fixed schema, and a parallel SQL query engine.
The release of data will be automatically distributed to different systems, where F1 supports a transaction-consistent replication across data centers. Thus, the overall system is said to cope with the failure of entire data centers without data loss. The promise of a high consistency, however, is associated with a disadvantage: F1 needs a little longer time than MySQL for writing data. The latency when writing to Google’s 50 to 100 milliseconds to read it be 5 to 10 milliseconds. The data throughput on the other hand will be significantly higher than that of MySQL. F1 has now replaced MySQL in Ad networks of Google.
The crap result of Google F1 reloaded
Those who have understood nothing, for them here is some useful links :
We have shown the real result previously how Google Ad Servers performs. Among the crap Alias names, ad.doubleclick.net, partner.googleadservices.com and pubads.g.doubleclick.net performs the worst when compared to leading Content Delivery Network (read Akamai).
What we want to say is quite clear, if Google blames MySQL for bad result to hide their quality of server for main business (that is Ad serving) and as it is measurable fact that the server hardwares are itself crap and the codes needs optimization, one can easily guess, with the name of providing good search result, what Google actually supplies.