Outcome
By the end of this lesson, you will be able to explain why prices differ between regions and weigh price against speed and rules when a region is chosen.
| Prerequisites | L1, I.1 L4 |
| Time | 7 minutes |
1. Concept
Every provider publishes a separate price for each region. The same machine, with the same size and the same software, can cost noticeably more in one region than in another.
Why prices differ
A provider’s own costs are not the same everywhere. Things that change from place to place include:
- Land and buildings. Space costs more in some cities than others.
- Electricity. Data centres use a lot of power, and power prices vary widely by country.
- Taxes and local rules. These change what it costs to operate in a place.
- Demand and scale. Large, busy regions spread fixed costs over more customers. Newer or smaller regions have fewer customers to share them.
The provider passes these differences on in its price list. It is the same reason a cup of coffee costs more in one city than in another.
Price is one factor, not the only one
From I.1 L4, you know that region choice also affects:
- Speed for users, which is better when the region is close to them.
- Rules about where data must be kept.
- Transfer costs, if the region is far from the other systems it talks to.
The cheapest region can be the wrong choice if it is far from customers or breaks a data rule. And a cheaper region can cost more overall if lots of data has to travel to and from it, as L3 showed.
When price is the deciding factor
Some work does not care where it runs. Examples: processing a batch of files overnight, training a model, or running tests. Nobody is waiting on a quick response, and the data may have no location rule. For work like that, picking a cheaper region is a reasonable way to save money.
Prices also change over time
Providers adjust prices, open new regions and introduce new machine types. A price you looked up a year ago may not be today’s. Always check the current pricing page before deciding.
2. Worked example
A team runs 10 identical machines all month, 730 hours each. It compares three regions. Prices are illustrative.
| Region | Illustrative price per hour | 10 machines x 730 hours | Monthly cost |
|---|---|---|---|
| Region A | $0.10 | 7,300 hours | $730 |
| Region B | $0.11 | 7,300 hours | $803 |
| Region C | $0.13 | 7,300 hours | $949 |
Moving from Region C to Region A would save $219 a month on these machines. But suppose the team’s customers are next to Region C and a contract says their data must stay there. Then Region A is not an option, and the extra cost is simply the price of serving those customers properly.
If instead the 10 machines run nightly reports that nobody watches live and that have no data rule, Region A is the sensible choice.
3. Hands-on (5 min: no account needed)
1. Open the Azure Pricing Calculator: azure.microsoft.com/pricing/calculator2. Add "Virtual Machines".3. Choose one region, a small Linux machine size, and 730 hours. Note the monthly cost.4. Without changing anything else, switch to two other regions. Note each cost.5. Work out the gap between the cheapest and the most expensive region.Then ask: if your customers were all in the most expensive of the three regions, would the saving be worth the slower response?
4. Knowledge check
Q1
Why can the same machine cost more in one region than another?
A. Providers charge more in regions that are new to them, always
B. The machine is faster in the expensive region
C. Prices are set at random for each region
D. The provider’s costs for land, power and tax differ by place
Show answer
Correct: D. Region prices reflect the provider’s local costs and demand. A is too strong; newer regions are not always the dearest. B is wrong because the machine size is the same. C is wrong because prices are published and deliberate.
Q2
Which job is the best candidate for running in a cheaper, more distant region?
A. A website that customers use in real time
B. Overnight batch processing with no data location rule
C. A system holding data that must stay in one country
D. A service that talks constantly to systems in another region
Show answer
Correct: B. Nobody is waiting on it and no rule ties it to a place. A needs to be close to users, C is tied to a country by rule, and D would add transfer costs that eat the saving.
Q3
10 machines each run 730 hours. One region costs $0.10 per hour and another $0.12 per hour (illustrative). What is the monthly difference?
A. $146
B. $73
C. $14.60
D. $1,460
Show answer
Correct: A. 10 x 730 = 7,300 hours. 7,300 x $0.02 = $146.
5. Apply
When you see the same service billed in several regions, compare their prices. If something costly runs in an expensive region, ask why it is there. “Our customers and our data rules are there” is a good answer. “That is where someone happened to create it” is a reason to look again.
Related lessons
- L3, Moving data: in is usually free, out is not
- I.1 L4: Where the computers live: regions and zones
- I.4: How teams actually use it (next)