Err from the prices given then the (non HK) Chinese should be complaining not Aussies.
Sadly I will never discover the joys of this thing. It turns out I am able to install and maintain my own database(s) and look after them.
Google's close to plugging a long-standing gap in its public cloud, with its Cloud Spanner distributed relational database hitting public beta. In January, we noted that Cloud Spanner, first detailed in a 2012 white paper, had landed as alpha in 2014 but was yet to become a commercial offering. The beta announced February 14 …
I can't wait until Amazon or Azure has got something like that.
I really feel that we should be able to have 'have cake and eat it' databases, and we should have had them for some time. Right now, this is a really cool feature which Google has but the others don't have. But even if it's brilliant, I'd feel a lot happier using it when Amazon or Azure have something that at least almost competes. I know that changing cloud provider would always be a nightmare, but I'd feel happier knowing that there is another cloud provider that does have roughly the right building blocks.
And just because I want to have my cake and eat it when it comes to databases (and cakes) doesn't mean I think this is applicable everywhere; notably it isn't a viable foreign policy.
Only caveat there is that AWS RDS gives you six database engines to choose from, not just MySQL. More flexible which explains my preference towards AWS as I've several tools to generate the code from models here. This is very useful tailoring codebase towards client requirements (i.e. what flavor are they capable of handling with their resources). Other than that, agreed.
Exactly. Seems like other commentors are not getting the "holy shit" of Google Spanner. This isn't another RDB. The banks have been after Google to release Spanner for years because they knew Google had it in their back pocket. The net of it is that RDBs are really good at data consistency and really terrible at scaling. NoSQL DBs are just the opposite, really good at scaling and do not have ACID properties like a traditional RDB so you can't run a transaction system on them, unless you don't care about your data being consistent. If you want to do something that requires crazy scale and also has ACID properties, you were smack out of luck and had to fragment the DB in some work around. Google has had an RDB which scales to PBs of data and thousands of nodes across a single instance (runs their ad auctions globally) with ACID properties of a traditional RDB like Oracle, DB2, etc.... how they get data consistency at that scale, I'm sure I don't know. Google magic. I plan to read up on it.... Anyway, that's why I wrote that if Google gets the word out, this is going to be huge. Nothing can do what Spanner does and everyone would like to do what Spanner does.
Spanner is Google's replacement for Hadoop for OLTP. The issue with Hadoop is that it has eventual consistency of data so it isn't suitable to a traditional RDB workload, like running an eCommerce service or the like. Spanner is the scale of Hadoop with the data consistency, ACID properties of Oracle. I would think Hadoop will be around because it does more than OLTP.... Oracle RAC, DB2 replacement is probably more of the use case... and any new workload of that sort. Huge relational DBs.