It’s smaller computers, doing limited tasks. So why is Edge Computing so complicated, and why are there so many definitions?
SHOW: 497
SHOW SPONSOR LINKS:
CLOUD NEWS OF THE WEEK - http://bit.ly/cloudcast-cnotw
CHECK OUT OUR NEW PODCAST - "CLOUDCAST BASICS"
SHOW NOTES:
THE COMPLEXITIES OF EDGE COMPUTING
No single definition of edge, because of so many use-cases
- Branch office
- Near-Edge
- Far-Edge
- Disconnected
Hardware can dictate quite a bit about edge characteristics
- Custom-OS or COTS OS
- General purpose HW or Specialized HW (GPU, NVMe, DPUs)
- #of computing devices; volume of local storage
Networking can dictate quite a bit about edge characteristics
- Are the devices always connected to the Internet?
- How are disconnected environments handled?
- How much bandwidth is available? Is it symmetric or asymmetric?
- What are the traffic patterns of the applications? Local vs. Backhaul.
Security can dictate quite a bit about edge characteristics
- How does Security happen on Day 0?
- What are the security vs. usability tradeoffs? (direct user interaction)
- How is security maintained? (CVE, Local Security services, etc.)
Maintenance can dictate quite a bit about edge characteristics
- How much does it cost to deploy new edges?
- How often are the edges expected to be upgraded?
- How will maintenance happen? (Local tech, remote download, etc.)
- How sophisticated is the Day 2 Ops? (Telemetry, Environmentals, Benchmarks, etc.)
FEEDBACK?
FEEDBACK?