- A server was rented but nobody watches it, so night-time incidents are handled by whoever notices.
- Backups exist but have never been restored as a test.
- The admin account sits with the previous provider, which makes moving away difficult.
- A company needs a GPU for an AI model but is not sure which card has enough memory.
Servers and GPU infrastructure
ALODEV owns its servers and GPUs. Its engineers monitor them, back them up and answer incidents. The same infrastructure runs our own software and the systems we build, and it is also offered as a service: application servers and database servers with monitoring, backup, security patching and incident response included.
Summary
- GPU lines
- H100, A100, RTX 5090 and RTX PRO 6000
- GPU memory
- 80 GB per H100 and A100 card, 96 GB per RTX PRO 6000 card, 32 GB per RTX 5090 card
- Servers
- From 1 to 12 vCPU, each a share of a real server, unlimited bandwidth
- Storage
- NVMe drives from 20 to 480 GB on every server plan
- Critical incidents
- Response time follows the maintenance plan in the contract
Problems we are asked to solve
What is included
| Server type | What it covers |
|---|---|
| Application server | Websites, apps and management systems, with the runtime environment, domain and security certificate set up. |
| Database server | PostgreSQL configured for you, backed up on a schedule, separated from the application server under heavy load. |
| GPU server | Training and running AI models and processing large data on H100, A100, RTX 5090 and RTX PRO 6000 cards. |
Every server plan includes the operations below.
- Critical incident duty: if the system stops or loses data, response time follows the maintenance plan in the contract.
- Backup and test restore: scheduled backups, restored as a test to confirm they work.
- Monitoring and patching: alerts on anything unusual, and operating system and software updates.
- Full access: admin accounts are handed over to your company, and can be moved elsewhere at any time.
GPU infrastructure
ALODEV has H100, A100, RTX 5090 and RTX PRO 6000 cards, used for training and running AI models and for processing large data. Memory decides what fits on a card: 80 GB on each H100 and A100, 96 GB on each RTX PRO 6000 and 32 GB on each RTX 5090. If you are planning an AI workload and are unsure which memory size it needs, we will help you work it out.
Ownership and moving away
Infrastructure is a service in its own right: you do not have to commission software from us to rent servers. If you later move to another provider, we hand over admin accounts, data and configuration, and we do not hold the keys. Where your data is placed is agreed before deployment, according to your needs and legal constraints; a domestic placement is available for data that must stay in Vietnam.
Frequently asked questions
Can we rent infrastructure without commissioning software?
Yes. Infrastructure is an independent service and does not require a software project.
How is this different from cheap server rental?
Cheap providers usually rent the machine only and leave operations to you. Here monitoring, security patching, backup and incident duty are included. If you already have people who run servers, renting a bare machine elsewhere can be cheaper.
Where is the data kept?
Agreed before deployment, according to your needs and legal constraints. A domestic placement is available for data that must stay in Vietnam.
What if we want to move to another provider?
We hand over admin accounts, data and configuration. We do not hold the keys.
Do you publish server prices?
Not on this page. Send us the workload and we quote in writing.
