Cloud infrastructure
Technical explanation for buyers: object, block and file storage, S3, Kubernetes, servers, latency and costs.
Infrastructure is a different product from Drive or iCloud. These guides explain the building blocks and what they cost, so you can compare technical choices.
Who is it for: Developers, devops and technical buyers.
Start here
- Step 1Object, block and file storage: differences and use casesObject, block and file storage solve different problems; the right choice depends on how your application accesses data.
- Step 2What is S3-compatible storage and what does it guarantee?S3-compatible means a service largely speaks the S3 API; which features really work is something you need to test yourself.
- Step 3What does object storage cost? Storage, requests and trafficObject storage costs more than just gigabytes: requests, outbound traffic and retention periods all count towards the total price.
All guides in this section
Cloud infrastructure
- S3 storage explained: what it is and how to chooseS3 storage explained: how object storage works, what the costs consist of, what to look for in security and how to choose an S3 provider.
- What is object storage (and S3)?Object storage and S3 explained for non-technical buyers.
- Object, block and file storage: differences and use casesObject, block and file storage solve different problems; the right choice depends on how your application accesses data.
- What is S3-compatible storage and what does it guarantee?S3-compatible means a service largely speaks the S3 API; which features really work is something you need to test yourself.
- What does object storage cost? Storage, requests and trafficObject storage costs more than just gigabytes: requests, outbound traffic and retention periods all count towards the total price.
- Block storage for databases and virtual machines explainedBlock storage gives a server a disk over the network; performance and recovery options determine whether it suits databases.
- Cloud storage for backups, archives and data lakesBackups, archives and data lakes use the same storage technology but call for a different set-up and different checkpoints.
- What is Kubernetes and when do you need it?Kubernetes automates running containers on multiple servers, but it also brings management effort that not every team needs.
- Managed Kubernetes or self-managed? How to weigh the choiceWith managed Kubernetes the provider runs the control plane; you remain responsible for workloads, updates and security.
- Managed Kubernetes or self-managed? How to weigh the choiceWith managed Kubernetes the provider runs the control plane; you remain responsible for workloads, updates and security.
- VPS, virtual machine, dedicated server and bare metal comparedFrom shared virtual server to your own physical machine: the choice comes down to isolation, predictable performance and management effort.
- Comparing S3 object storage providersComparing S3-compatible object storage on price per GB, data transfer costs, regions and lock-in.
- What are egress costs in the cloud and how do you limit them?Egress is outgoing data traffic; it can become an unexpectedly large cost item with many downloads, migrations or traffic between regions.
- Cloud latency explained: how region and network affect performanceLatency is the delay in a connection; distance to the region and the number of network hops determine how fast a service feels.
- Cloud latency explained: how region and network affect performanceLatency is the delay in a connection; distance to the region and the number of network hops determine how fast a service feels.
- Edge computing or cloud computing: what is the difference?Edge processes data close to the source, cloud centrally; many applications combine both for speed and scale.
- High availability in the cloud: what it means for architectureHigh availability is about absorbing failures with redundancy, zones and automatic failover; costs and complexity increase with ambition.
- High availability in the cloud: what it means for architectureHigh availability is about absorbing failures with redundancy, zones and automatic failover; costs and complexity increase with ambition.
- Designing a cloud architecture for peak loadAn architecture for peaks scales automatically, decouples components and protects itself; testing before the peak is essential.
- What cloud infrastructure do you need for AI?Which components an AI application needs: computing power, storage, network and management, and how to link them to your goal.
- Comparing GPU cloud: what to look for with AI workloadsCriteria for comparing GPU cloud offerings: memory, network, availability, storage, region and terms.
- Comparing GPU cloud: what to look for with AI workloadsCriteria for comparing GPU cloud offerings: memory, network, availability, storage, region and terms.
- Renting or buying GPUs for machine learning?Rent or buy: what suits varying or constant GPU use, and which costs and risks come with it.
- AI training, inference and fine-tuning: what do they need?The difference between training, fine-tuning and inference, and what each requires in compute, memory, storage and networking.
- AI training, inference and fine-tuning: what do they need?The difference between training, fine-tuning and inference, and what each requires in compute, memory, storage and networking.
- Private AI or public AI: consequences for data and infrastructurePublic AI service or private AI in your own environment: what does it mean for data, control, costs and management?
- Private AI or public AI: consequences for data and infrastructurePublic AI service or private AI in your own environment: what does it mean for data, control, costs and management?
- Data storage for AI: object storage, data lakes and datasetsWhich storage suits AI: object storage for datasets, fast storage for training and a good structure for your data.
- Hosting AI in production: from prototype to reliable serviceWhat changes when you take an AI model from prototype to production: capacity, monitoring, security and recovery.
- Planning GPU capacity for peak load and experimentsHow to plan GPU capacity for steady workloads, peaks and experiments without overpaying.
- Planning GPU capacity for peak load and experimentsHow to plan GPU capacity for steady workloads, peaks and experiments without overpaying.
- Calculating AI infrastructure costs: GPUs, storage and networkA calculation model for AI costs: GPU hours, storage, network traffic, management and the costs that are often forgotten.
- AI data privacy: which infrastructure choices help?Which infrastructure choices help protect privacy with AI: region, encryption, access, logging and contracts.
- What is a GPU and what do you use one for in the cloud?An explanation of the GPU: how it works, what you use it for in the cloud and when you do not need one.
- What is a CPU and what does vCPU mean for cloud servers?An explanation of CPU, cores, threads and vCPUs on cloud servers and how to choose what you need.
- CPU or GPU: which computing power suits your cloud work?How to choose between CPU and GPU for your cloud work, with examples per type of workload.
- What is VRAM and how much GPU memory do you need?Explanation of VRAM, what it determines and how to estimate how much GPU memory your AI work needs.
- GPU in the cloud for rendering, video and simulationsHow to use GPU capacity in the cloud for rendering, video and simulations, and what to look out for.
- Choosing a GPU instance: a checklist for your providerA checklist of questions for your provider before you choose a GPU instance in the cloud.
Frequently asked questions
What is the difference between object, block and file storage?
Object storage keeps individual objects, block storage works like a disk and file storage shares folders over the network. The guide compares the three.
When do I need Kubernetes?
Only when you have several services that need to scale and recover automatically. For a simple application it is often too heavy.
How do I compare object storage costs?
Include storage, requests and outgoing data traffic. The guide shows the calculation model without amounts.
Other sections
- Section 1General cloud guidesThe basics: what cloud storage is, which types of cloud exist and how to choose a cloud provider.
- Section 2Personal: photos, files and backupCloud storage for home: photos, documents, phone backup and privacy.
- Section 3Backup and recoveryBacking up your phone and computer, testing, restoring and dealing with ransomware.
- Section 4NAS and private cloudYour own storage at home or in the office: what a NAS is, how to arrange backups and when a private cloud fits.
- Section 5Start-ups, freelancers and SMEsCloud for small teams: sharing files, choosing a cloud provider, keeping costs manageable and migrating.
- Section 6Enterprise, IT and complianceCloud choices for large organisations: governance, compliance, recovery, costs and an exit strategy.
Last updated on
