Cloud Hosting Features Explained for Beginners: What Each One Does and Costs
A beginner's glossary of cloud hosting features, including instances, snapshots, load balancers, autoscaling, managed databases and CDNs, with cost notes.
A cloud console can look like a menu with hundreds of items, but a small business only needs to understand about a dozen features. Virtual servers run your software, storage keeps your data, snapshots and backups protect it, load balancers and autoscaling provide resilience, managed databases remove maintenance, and a CDN speeds delivery and cuts transfer costs. This guide explains each feature in plain language, what problem it solves, how it is usually billed and when you can skip it.
Compute: the servers that run your software
Virtual machine (instance, droplet, VM). A virtual computer with a chosen number of virtual CPUs (vCPUs), memory and disk. Sizes are usually named small, medium, large and so on, and billed hourly or monthly.
Container service. Runs software packaged in containers, so it behaves the same everywhere. It is useful if your developers already use containers, and it often scales more easily than hand-built servers.
Serverless functions. Small pieces of code that run only when triggered and are billed per execution and runtime. They suit scheduled tasks, webhooks and light APIs. They can cost more than a server for constant heavy use.
Platform services. You supply code and the platform handles servers. Convenient, but cost per unit of capacity is usually higher.
Storage: where data lives
Block storage. Virtual disks attached to a server, billed per gigabyte per month. Good for operating systems and databases.
Object storage. A place for files such as images, documents and backups, accessed over the internet through an interface. It is cheaper per gigabyte than block storage and scales almost without limit, but it is billed for requests and transfer too.
File storage. A shared network drive that several servers can use at once. It is more expensive per gigabyte than object storage.
Protection: snapshots and backups
Snapshot. A point-in-time copy of a disk. Fast to take, billed by stored size, and useful before risky changes.
Backup. A managed schedule of copies with retention rules. Providers may charge a percentage of the server price or by gigabyte. For a safe plan, keep copies away from the system they protect. Our backup guide covers the details.
Networking: connecting and protecting
Public IP address. The address on the internet where your server can be reached. Some providers charge for it hourly.
Firewall or security group. Rules about what traffic can reach your servers. Free, and essential.
Load balancer. Spreads incoming requests across several servers and removes unhealthy ones. Billed hourly plus usage. It provides resilience and enables updates without downtime.
DNS service. Translates your domain name into addresses. Often free or inexpensive.
Virtual private network or private network. Lets servers talk to each other without exposing traffic to the internet.
CDN (content delivery network). A global network that caches copies of your files near users. It speeds up delivery and reduces load and transfer from your servers. Billed per gigabyte and request.
Scaling: handling more or less traffic
Vertical scaling. Moving to a larger server. Simple, but usually requires a restart and has a ceiling.
Horizontal scaling. Adding more servers. It requires software that can run on more than one machine, with shared storage or a shared database.
Autoscaling. Rules that add or remove servers automatically as load changes. It saves money on variable workloads, but requires careful setup and limits so a traffic spike or an attack does not cause an enormous bill.
Data services
Managed database. The provider operates a database such as MySQL, PostgreSQL or others, handling installation, patches, backups and failover. It costs more than running the database yourself on a server but saves effort and reduces risk. Billed by instance size, storage and backup retention, with extra for replicas.
Caches and queues. Services that speed up applications or let components communicate reliably. Add them when you have a real need, not by default.
Operations: seeing and controlling
Monitoring and alerts. Collects metrics like CPU and memory, and notifies you of problems. Basic monitoring is often included, with detailed metrics billed.
Logging. Stores application and system logs, billed by volume ingested and stored.
Identity and access management. Controls who can do what in your account. Use it to avoid shared logins.
Cost management tools. Budgets, alerts and cost reports. They are usually free and should be turned on first. See our monitoring and cost alerts guide.
Which features do you need?
| Feature | Skip if | Add when |
|---|---|---|
| Load balancer | One server and downtime is acceptable | Downtime costs more than the balancer's price |
| Managed database | You have in-house database skills and a small load | You lack time to manage backups and updates |
| Autoscaling | Traffic is steady | Traffic varies widely |
| CDN | Site is small and local | You serve images, video or global users |
| Multi-region | Single-country customers and tolerable risk | Regulatory or very high availability needs |
| Serverless | Constant heavy workload | Occasional or event-driven tasks |
Worked example: price of resilience
This is an illustrative example. One server at 40 dollars a month runs a customer portal. Adding a second server (40 dollars), a load balancer (18 dollars) and upgrading the database to a high-availability pair (an extra 60 dollars) lifts the bill by 118 dollars to 158 dollars a month. If the portal handles 4,000 dollars of billing per day, an extra day of downtime avoided each year more than covers the upgrade's annual cost of 118 x 12 = 1,416 dollars. Make this calculation with your own numbers.
Terms that sound similar
- Availability zone: a separate data center or group within a region, for resilience.
- Region: a geographic area containing several zones.
- SLA: the provider's uptime promise and the remedy, usually credits, if it is missed.
- Uptime percent: 99.9 percent allows roughly 8.8 hours of downtime a year, and 99.99 percent allows about 53 minutes.
Mistakes beginners make
- Turning on every feature a tutorial mentions.
- Using the largest server "to be safe."
- Leaving databases open to the whole internet.
- Forgetting that stopped servers may still bill for storage.
When to consult a professional
Get help when designing redundancy for a critical system, handling regulated data, or setting up identity and network controls. A short consulting engagement is cheaper than recovering from a design mistake.
Frequently asked questions
What is the difference between a snapshot and a backup?
A snapshot is a copy of one disk at one time. A backup is usually a scheduled, retained, restorable copy managed by a policy. Use both.
Do I need Kubernetes?
Most small businesses do not. It adds complexity and is worthwhile mainly when you run many services.
Is serverless cheaper?
For infrequent workloads, often yes. For constant heavy traffic, servers can cost less.
What is infrastructure as code?
Writing your setup in files that tools can apply, so environments can be rebuilt consistently.
Sources
- National Institute of Standards and Technology — cloud computing definitions
- Cybersecurity and Infrastructure Security Agency — security guidance for organizations