Talk by Arun Nagarajan at Google Developer Group London.
Cost Reduction through platforms
1957 - 2013: steep decrease in average S&P 500 company age
Easier to build products:
- affordable capacity
- on demand computing
- instant access
Google handles lots of data
- 72 hours of videos uploaded to Youtube every minute
- 100 million GB - size of search index in 2010
- 425 million active Gmail users
Cloud Computing
As a Service:
- IaaS: Infrastructure, e.g. more choice than PaaS
- PaaS: Platform, e.g. Google App Engine
- SaaS: Software, e.g. Jira, Github
IaaS vs PaaS: IaaS means controlling Runtime, Middleware and OS in addition to Applications and Data (as in PaaS).
Google Cloud Platform
Compute
- App Engine
- Compute Engine Storage
- Cloud Storage (blobs of files)
- Cloud SQL (RDBMS)
- Cloud Datastore (NoSQL) App Services
- BigQuery
- Cloud Endpoints
Compute Engine (IaaS)
“Think of it as a general-purpose Linux machine in the cloud.”
Compute Engine Setup
- console.developers.google.com
- Compute Engine requires billing to be enabled.
- Need to install Google Cloud SDK.
- Create a Google Cloud instance.
- => This creates a Linux box that you can SSH in and run your apps.
- Need to add firewall rule to allow access from port 80.
- Need to install packages yourself with apt-get (e.g. Python, Node)
- Can install arbitrary libraries/frameworks
Managed VMs
- Allow you to build app engine modules that use Compute Engine