At first glance IaaS providers look alike: CPU, memory, disk and a price. The real differences show up in details such as data centre location, KVKK compliance, SLA, backup, network security, support model and exit strategy. We share 10 criteria and a checklist to help you choose the right provider.
You have decided to move your servers to the cloud. When you put the quotes side by side, they all look similar: so many vCPUs, so much RAM, so much disk and a monthly amount. But a few months later, what decides satisfaction or disappointment is rarely those lines.
IaaS (Infrastructure as a Service) means getting virtual servers, storage and network infrastructure as a service. The operating system and applications stay under your control. If you are not yet sure whether IaaS or PaaS fits you better, settle that first. This article covers the ten questions a business that has chosen IaaS should ask when picking a provider.
1. Data Centre Location and Infrastructure Level
Your virtual server physically runs somewhere. Where that place is, and how it is built, directly affects both performance and resilience.
- Location: a data centre close to your users means lower latency. For users in Turkey, an in-country data centre responds noticeably faster than one abroad.
- Infrastructure level (Tier): this describes redundancy in power, cooling and network connectivity. Ask the provider which data centres they operate in and what redundancy those facilities have.
- Multiple locations: can you keep a copy in a second data centre for disaster recovery?
2. Data Residency and KVKK
If your servers process personal data such as customer, employee or patient records, the country where that data is kept is a legal question. Transferring personal data abroad is subject to additional conditions under KVKK, Turkey's data protection law.
Keeping data in Turkey removes the need for those extra processes, although it does not end your other KVKK obligations. Ask the provider where not only the servers but also the backups, logs and disaster recovery copies are kept.
3. SLA: What Is Guaranteed, and How Is It Measured?
A service level agreement (SLA) is the provider's availability commitment. Looking only at the percentage is not enough:
- Which components does the SLA cover: virtual servers, storage, network, all of them?
- Is planned maintenance counted as downtime?
- What happens if the commitment is missed: service credits or another remedy?
- Are first response and resolution times defined for incidents?
Just as important as the percentage on paper is who steps in, and how fast, when something goes wrong.
4. Virtualisation Platform
The provider's virtualisation platform (VMware vSphere, KVM-based solutions, Hyper-V and so on) affects ease of migration, supported operating systems and backup options.
If your current environment runs on VMware and you move to a provider on the same platform, migrating virtual machines usually requires less conversion. If you are moving to a different platform, discuss in advance how driver and disk format conversions will be handled. Recent licensing changes in the virtualisation market are also worth factoring into long-term cost.
5. Are Backup and Disaster Recovery Included?
This is the most common wrong assumption: "it is in the cloud, so it must be backed up". In most IaaS services the infrastructure is redundant, meaning a disk or host failure will not affect you. But that redundancy does nothing for an accidentally deleted file, a corrupted database or ransomware.
- Is virtual server backup included, or charged separately?
- How many days are backups kept, and are they stored in a separate repository?
- Is there an immutable backup option?
- Who performs restores, and how long do they take?
- Is disaster recovery with replication to a second location possible?
6. Network, DDoS Protection and Public IPs
If your servers are exposed to the internet, the network layer matters at least as much as compute.
- Bandwidth: what is the port speed, and is there a traffic quota?
- DDoS protection: is network-level protection standard? If application-layer (L7) protection is needed, how is it provided?
- Public IPs: how many public IP addresses are included, and how do you get more?
- Firewall and VPN: are virtual firewalls, site-to-site VPN to your offices and private networks (VLANs) available?
7. Support Model: A Direct Engineer or a Ticket Queue?
When something breaks, who will you be talking to? Anyone can buy infrastructure from the global hyperscalers, but support there is usually tied to paid tiers and runs through tickets with teams who do not know your systems.
With a managed service provider, you can talk directly to the engineers who actually operate the systems. Questions to ask:
- In which language, and during which hours, is support provided?
- Is there a named engineer who knows your environment?
- Are management tasks such as operating system updates, monitoring and backup checks included?
- Who do you reach in an emergency, and how?
8. Pricing Transparency and Egress
Most cloud bill surprises come from items that were not visible in the first quote. In usage-based models, data egress, disk operations, snapshots, public IPs and support can all be billed separately.
Fixed monthly billing makes budgeting easier. When reviewing a quote, ask: "What is included in this amount, and when do extra charges apply?" Also clarify which currency the price is in and under what conditions it is updated.
9. Migration Support
Who will move your existing servers to the new environment? Some providers simply hand over an empty virtual machine and leave the migration entirely to you. Others run the planning, testing and cut-over night with you.
- Is there a pre-migration inventory and sizing exercise?
- Is there a test migration and a rollback plan?
- How is the downtime window planned?
- Is the migration fee included in the quote?
10. Exit Strategy
Thinking about the exit when you sign up may feel odd, but it is a sign of a healthy relationship. When the contract ends or you want to move to another provider:
- Can you export your virtual machines in a standard format?
- How is data handed back to you and then securely deleted?
- What are the contract term and termination conditions?
- Are there extra charges for bulk data egress?
A Practical Way to Compare Quotes
Knowing the ten criteria is one thing; making quotes genuinely comparable is another. Providers package items in different ways, so send every provider the same definition of your needs.
- Build an inventory: number of servers, CPU, memory and disk usage for each, operating system and the applications running on it.
- State criticality: how many hours of downtime can each system tolerate, and how much data loss is acceptable?
- Write down management expectations: who will handle updates, monitoring and backup checks?
- Calculate total cost: put servers, backup, licences, traffic, support and migration in a single table.
A request prepared this way leads to more accurate quotes and fewer "that was not included" surprises after migration.
Checklist
You can use the table below when comparing quotes:
| Criterion | Question to Ask | A Good Answer |
|---|---|---|
| Data centre | Which facility, what redundancy level? | A named facility with redundant power and network |
| Data residency | Where are servers, backups and logs? | All in Turkey |
| SLA | What is the scope, measurement and remedy? | Written, measurable, with defined response times |
| Platform | Which virtualisation platform? | Clearly stated and compatible with your environment |
| Backup | Included? Immutable? | Separate repository with an immutable option |
| Network and DDoS | Quota, protection, public IPs? | Unmetered traffic, standard network-level protection |
| Support | Who will I talk to, in which language? | A Turkish-speaking engineer who knows the environment |
| Pricing | Which items trigger extra charges? | Fixed monthly, with clear line items |
| Migration | Who performs the move? | A planned, tested migration run together |
| Exit | How do I get my data back? | Export in standard formats, written deletion process |
Frequently Asked Questions
What advantages do IaaS providers in Turkey offer over global providers?
The clearest advantages are keeping data in Turkey, low latency for local users, simple fixed monthly billing and working directly with a Turkish-speaking engineer who knows your environment.
Is it right to decide on price alone?
No. A quote that looks cheap can end up more expensive once backup, support, traffic and migration are added. Compare quotes on the same scope.
Does moving to IaaS hand all server management to the provider?
In pure IaaS, the operating system and applications are your responsibility. With managed IaaS, tasks such as updates, monitoring and backup checks can be taken on by the provider. Make the split of responsibilities clear in the contract.
How long does migration take?
It depends on the number of servers, data volume and application dependencies. An environment of a few servers can usually be moved in a planned weekend cut-over; larger environments are migrated in phases.
Keep Your Servers in Turkey with CloudRUX IaaS
CloudRUX IaaS runs your virtual servers in data centres in Turkey, with fixed monthly billing and no surprise line items. Our real difference is that you work with an engineer who actually operates the systems and knows your environment. See the details on our IaaS page, or contact us for a tailored quote and migration plan.