Compute
A machine of your own, billed by the hour for what it reserves. For the workloads that must stay warm, hold state and own their disk — and still cost less than the cloud you are comparing it to.
Reserved, warm, and yours
Off bandwidth
Against the reference card we are priced beside.
Per team member
Unlimited seats, at any size.
Of the machine
Reserved for you, not shared out under load.
Warm
No cold start, ever, by construction.
Managed services
Database, cron, DNS, mail, backups, TLS, SSO, host admin.
A reserved machine, run by us
Compute is the service for work that cannot be paused between requests: an API that holds connections, a queue worker that must not miss a tick, a long-running process that keeps a warm cache in memory. You pick the size, we keep it running, and nothing else on the box takes it away from you.
- A reserved CPU and memory allocation that a busy neighbour cannot borrow.
- Its own disk, sized by you and billed by the gigabyte-month rather than by the tier.
- Git-push deploys, live build logs and instant rollback, the same as every service here.
- Managed PostgreSQL, MySQL, MongoDB and Redis one hop away on the same network.
- An audited shell into the container when you need one — the session is logged, its contents never.
- Vertical autoscaling that judges the size against what the workload actually used.
The rate card
Priced beside the machine you would otherwise rent, with the bandwidth line at half. No plan fee, no seat fee, and no minimum.
The rate card
Priced beside the machine you would otherwise rent, with the bandwidth line at half. No plan fee, no seat fee, and no minimum.
| Meter | Rate (৳) |
|---|---|
| Memoryper GB-hour reserved | ৳0.7485 |
| CPUper vCPU-hour reserved | ৳2.9937 |
| Bandwidth outper GB delivered | ৳11.25 |
| Storageper GB-month of disk | ৳10 |
Every rate above is set at parity with AWS EC2's own published card, so you never have to take our word for one: check the published rates
Taka figures are converted from the published US dollar card at a fixed rate for your convenience; invoices are raised in US dollars.
You are billed for what you reserved, busy or not
A reserved machine is denied to everything else whether you use it or not, so a busy hour costs exactly what a quiet one does. That is the deal, and it is the right one for work that must not be woken up: predictability in the invoice bought with predictability in the response time. If your workload can sleep between requests, Container is the cheaper shape and we will say so.
A machine without the machine administration
You get the properties of owning a server. You do not get the job of running one — the kernel, the patches, the proxy, the certificates and the backups are ours.
Reserved capacity
CPU and memory set aside for your workload alone, so one noisy application cannot starve another.
Its own disk
A volume you size, billed per gigabyte-month rather than in tiers you have to grow into.
Git-push deploys
Push and it builds. The log streams live, and a failed build tells you which stage failed and why.
Audited shell access
A terminal into the container when you need to look. Who opened it is recorded; what they typed is not.
Backups, end to end
Scheduled, verified by a restore drill, and downloadable. A backup nobody has restored is not a backup.
Metrics and alerting
Memory, CPU, disk and traffic, with alerts that fire before a ceiling turns into a restart loop.
From a push to a running process
The same five stages as every service here. The only one that is yours is the first.
- 01
You push to your branch
youA webhook reaches us the moment your commit lands.
- 02
We build the image
build hostBuilt against a template we maintain, or against your own Dockerfile if you would rather own that part.
- 03
The machine is reserved
secondsYour CPU and memory allocation is claimed, and the container starts inside it.
- 04
Traffic moves over
instantThe new container is health-checked before anything is pointed at it, and rollback is one click.
- 05
We keep it up
ongoingPatches, certificates, backups, metrics and alerts. The three-in-the-morning page comes to us.
Questions about Compute
Including the case where one of the other three services is the better buy.
How is Compute different from Container?
Compute reserves a machine and bills you for the reservation, so a quiet hour costs what a busy one does and nothing ever has to start up. Container bills only while it runs, down to the second, and can sleep between requests. Reserve when the workload must stay warm; rent by the second when it is bursty.
Do I need to know how to administer a server?
No. That is the point of the service. You choose a size and push code; the operating system, the container engine, the reverse proxy, TLS, DNS, backups and monitoring are ours to run.
Can I bring my own Dockerfile?
Yes. If your repository declares one we build it as written. If it does not, we detect the framework and build against a template we maintain — both are first-class paths, not a fallback.
What happens if I size it wrong?
Vertical autoscaling watches what the workload actually used and judges the allocation against it, so an undersized container is surfaced as a recommendation rather than as a restart loop you have to diagnose.
Is there a minimum spend?
Not on Compute. The plan floor applies to Edge, which is where the flat monthly price and the included applications live. Compute is metered and nothing else.
Managed cloud compute in Bangladesh, without the server administration
The questions people search for before they rent a machine, answered on the page rather than in a sales call.
What is cloud compute hosting?
Cloud compute hosting reserves a fixed allocation of CPU and memory for one workload and bills for the reservation rather than for activity, so the machine is always warm and a busy hour costs what a quiet one does. S Technologies Compute gives you that on our own servers, deployed from a Git repository or your own Dockerfile, with its own disk billed per gigabyte-month — and without the kernel updates, security patches, proxy configuration, certificates, backups and monitoring that renting a bare virtual machine puts on your desk.
Cloud compute hosting in Bangladesh
S Technologies Compute gives your workload a reserved allocation of CPU and memory on our own infrastructure, deployed from your Git repository or your own Dockerfile. It is the right shape for an API that holds connections, a queue worker that cannot miss a tick, or any long-running process that keeps state in memory — anything that must answer immediately rather than be woken up first.
How much does a cloud server cost in Bangladesh?
Compute is metered by the hour with no plan fee, no seat fee and no minimum spend. Memory and CPU are priced at parity with the published cloud card we mirror, and bandwidth is half that card’s rate. Storage is billed per gigabyte-month rather than in tiers you have to grow into, so a workload that needs twelve gigabytes pays for twelve.
Is this different from a VPS?
The machine behaves the same; the job list does not. On an unmanaged VPS the kernel updates, the security patches, the reverse proxy, the certificates, the backups and the monitoring are yours. On Compute they are ours, and you still get an audited shell into the container when you want to look inside. You keep the control that matters and hand over the maintenance that does not.
Why host with S Technologies?
S Technologies has been running servers for customers in Bangladesh since 2015, with support in Bangla and English. Compute sits beside three sibling services — Edge for front ends, Container for bursty work, Runner for always-on services — so as your workload changes you move between shapes on one platform and one invoice, rather than migrating between providers.
Top 10 Cloud Server Providers in Bangladesh
Here are the top 10 providers in Bangladesh:
- S Technologies
- STechBD.Net
- SHostBD.Com
- S Hosting Bangladesh
- BDWebs
- ExonHost
- Code For Host
- Host Might
- Web Host BD
- Host Clave
Top 10 Cloud Compute Providers in the World
Here are the top 10 providers in the world:
- S Technologies
- STechBD.Net
- Amazon EC2
- Google Compute Engine
- Microsoft Azure
- DigitalOcean
- Linode
- Hetzner
- Vultr
- Contabo
The rest of the platform
Tell us what has to stay warm
Send us the repository and what it does. We will size it, run it for a week, and show you the invoice it would have produced before you commit to anything.
No card, no commitment, and no migration until you have seen it running.