Cloud Storage is one of those services on the Professional Cloud Architect exam where the wrong default can cost a customer real money. Picking the right storage class and the right location is mostly about matching how often the data gets read against how long it needs to live somewhere, and then layering on durability and compliance. I want to walk through the four storage classes and the three location options the way they actually appear on the exam.
Google Cloud Storage has four classes: Standard, Nearline, Coldline, and Archive. They differ on two axes that matter for the exam: storage cost per gigabyte, and minimum storage duration. The cheaper the storage class, the longer Google requires you to keep the data there before deleting it without penalty.
Standard has no minimum storage duration. It is the most expensive class per gigabyte, but it is built for data that needs to be served with low latency and accessed frequently. Live application data, user-uploaded files behind a website, anything that gets read every day, this is what Standard is for.
Nearline has a 30-day minimum storage duration. It is meant for data accessed less than once a month. The classic example is monthly reporting data or backups you only restore from in an emergency. You still get fast access when you need it, but you are charged a retrieval fee on top of the lower storage cost.
Coldline has a 90-day minimum. It is for data accessed once every few months at most. Disaster recovery archives are the canonical fit. The data has to be available, but you are not pulling from it on any regular cadence.
Archive has a 365-day minimum. It is the cheapest storage class and is designed for data accessed once a year or less. Compliance archives, audit logs you have to retain for seven years, long-term records that almost certainly will never be read again, all of these go in Archive.
The minimum duration is what trips people up. If you put data in Coldline and delete it after 30 days, you still pay for the full 90 days. So the question is not just "how often will I read this" but "am I confident this data will live here for at least the minimum period." Choosing the wrong class up front is the most common mistake.
There is one nuance worth knowing for the Professional Cloud Architect exam. Archive is recommended for data accessed once a year or less, but it can still be the right choice for data accessed two or three times a year if cost savings is the priority. The reasoning is that Archive's storage cost is so much lower than Coldline's that the savings on storage can outweigh the higher retrieval fees, even if you pull the data a few times.
This only works if you have done the math. The storage savings need to be greater than the cumulative retrieval costs. For a small dataset accessed three times a year, Coldline might still win. For a large dataset accessed three times a year, Archive often comes out ahead. The exam expects you to recognize that Archive is not strictly limited to once-a-year access patterns when cost is the dominant constraint.
You set a default storage class on the bucket, and every object you upload inherits it. But you can override the storage class on individual objects within the bucket. This is useful when most of your data follows one access pattern but a subset needs different handling. You can also use Object Lifecycle Management to transition objects between classes automatically based on age, which is a more practical way to manage this at scale than overriding objects one at a time.
Storage class controls cost based on access frequency. Location controls where the data physically lives, and that drives availability, latency, and a different cost dimension.
Regional storage keeps your data in a single region. This is the right choice when compliance requires the data stay in a specific geography, when your application is also in that region and you want low-latency reads, or when cost matters and you do not need the durability guarantees of a broader footprint. Regional is the cheapest of the three location types.
Multi-region storage replicates your data across a continent. The three multi-region locations are US, EU, and Asia. Multi-region gives you the highest availability and lets you serve data with low latency from anywhere within that continent. The trade-off is cost. You are paying for storage replicated across multiple regions, so the per-gigabyte rate is higher than regional. Use it when the workload genuinely needs continent-scale access or when availability requirements are strict enough to justify the spend.
Dual-region sits between regional and multi-region. The data lives in two specific regions you choose, which gives you higher durability and availability than a single region while being cheaper than full multi-region replication. The classic use case is disaster recovery. You get a geographically distributed copy of the data without paying for replication across an entire continent.
Most Professional Cloud Architect questions about Cloud Storage give you a scenario and ask you to pick the cheapest option that meets the requirements. Read the access pattern carefully. If the data is accessed daily, it has to be Standard. If it is accessed monthly, Nearline. Quarterly, Coldline. Yearly or less, Archive, with the caveat above for cost-driven scenarios.
For location, look for compliance constraints first. If the data must stay in a country or region, that forces regional. If the requirement is global access or high availability across a continent, multi-region. If the requirement is disaster recovery with two specific regions, dual-region. Cost-sensitive scenarios with no specific availability requirement default to regional.
The exam will often combine these. A question might describe a healthcare archive that must stay in the US, is accessed twice a year, and needs to be retained for seven years. That is regional Archive. Another might describe a global media library that needs to serve users across Europe with low latency. That is multi-region Standard in EU. Reading the access frequency and location constraints carefully gets you to the right answer almost every time.
My Professional Cloud Architect course covers Cloud Storage classes and location options alongside the rest of the storage and analytics material.
