How Google Spanner's easing our distributed SQL database woes I've been messing about with databases for a long time. I say "messing about" because I've never been a DBA, but as a systems programmer and storage administrator, I've been on the periphery of the application layer and of course I've deployed many personal databases. I was intrigued to read about the new global distributed …
The only thing even like Spanner is the open source CockroachDB. It was written based on the Google Spanner paper. It is Spanner but without TrueTime which is the Direct fiber, atomic clocks and GPS but critical it does not have the Google algorithms for guarantee maximum drift.
What is here that will be very difficult to replicated is the direct fiber and private network that Google owns that crosses the globe.
We can now see why Google invested the billions recently to lay fiber below the pacific ocean creating the fastest inter-pacific connection. This is key as you can NOT do this over public IP and you MUST have predictable network latency.
Each data center has a cesium clock for accuracy and its synced to the National Atomic clocks receiving a radio time signal. Based on the known source of the clock and their GPS location you can accurately sync and keep time in each data center. Pretty simple and the hardware is approximately 4-6K (USD) per data center.
Even if your GPS resolution is to within 100 meters, your clock is more accurate than the clock in a server.
So Google isn't doing anything fancy here.
The rest of the technology isn't that difficult either.
The larger issue is the legalities that will vex Google.
When spanner crosses a country's border, which country's data laws are in effect?
Think about that for a second.
That's the stuff that will kill Google. And there's already two cases. One against Google and One against Microsoft on US wanting access to data stored outside of the US.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
