Cloud providers tout their data center count – Microsoft 12, SoftLayer 9, Amazon 8. But why does it matter? David Mytton explains location importance.
Translation:
The cloud infrastructure market is fiercely competitive. Building infrastructure requires massive capital, making location critical.
Why does location matter?
Diverse locations matter for:
- Redundancy:While rare, data center outages happen. Distributing load across facilities prevents single points of failure.
- Data protection:Different data types have different location requirements (e.g., EU data must stay in EU).
- User latency:For latency-sensitive apps, data centers should be close to users. CDN helps but connectivity is essential.
Worldwide data centers aren’t cheap. Large providers have advantages in scale and efficiency.
Current situation
Different providers have regions. Internal redundancy isn’t enough – entire regions can fail.
Data Center Locations by Geography
|
Provider |
USA |
Europe |
Asia |
Other |
|
Amazon Web Services |
3 |
1 |
3 |
1 |
|
Microsoft Azure |
5 |
2 |
4 |
1 |
|
|
1 |
1 |
1 |
0 |
|
Rackspace |
3 |
1 |
2 |
0 |
|
Softlayer |
5 |
2 |
2 |
0 |
Azure leads with 12 regions, followed by SoftLayer 9, Amazon 8, Rackspace 6. Google lags with 3.
Major player locations
Amazon long had only one European data center. For redundancy, you need at least 2 close together to avoid latency.
SoftLayer is expanding with $1.2B in new data centers in Hong Kong, London, and more locations planned.
Google spends heavily on infrastructure but many data centers aren’t part of Google Cloud yet. As a newcomer, it needs to catch up.
About China
China is an interesting case. Cross-border traffic increases latency. Microsoft and Amazon have China data centers requiring separate accounts. Google left China 4 years ago.
Location is a competitive advantage. Microsoft leads now but may lose to SoftLayer. Cloud availability will keep expanding.
Original link: The next big front for cloud competition: Location, location, location（翻译/海霞 责编/仲浩）