What determines a good VDS purchase is not the amount of RAM but the answer to four questions: is your workload sensitive to single-thread speed or to core count, where are your visitors located, who is going to administer the server, and in which country does your data have to live? Buying without settling those four leaves you either with an oversized server sitting idle or with an undersized plan you will have to migrate off within three months. Below is what to check before you buy, how to size the configuration properly, and which clauses actually matter in the contract.
The most common mistake in server buying is opening the price list and picking "something in the middle". The correct order is the reverse: define the workload, then look for the configuration that fits it.
Ask yourself:
With those answers in hand, choosing a plan becomes mechanical.
This is the most misunderstood part. "More cores is better" is not true for every workload.
Single-thread sensitive workloads — Minecraft and other game servers, classic PHP applications, most WordPress installs — do their work on one core and benefit visibly from a high-clocked processor. On this kind of workload, an eight-core previous-generation server can be slower than a four-core current-generation one.
Multi-core workloads — hosting many sites, running containers, video transcoding, parallel builds — scale close to linearly with core count.
When comparing offers, ask for the processor generation and clock speed. The AMD Ryzen 9 9950X, for example, is a Zen 5 part with 16 cores and 32 threads, a 4.3 GHz base clock and up to 5.7 GHz boost. A VDS partitioned from a processor like that is strong on both single-thread and multi-core work.
RAM is not the place to economise, because running out means either a crash or a slide into swap and severe slowdown. A rough starting scale:
If utilisation sits consistently above 80%, it is time to size up.
The first storage question is not "how many GB" but "which technology". NVMe M.2 SSD delivers multiples of the IOPS a SATA SSD does, and the gap is felt most in database-heavy applications. On capacity, total up the operating system, application, database, log files and local backups — then add at least 30% headroom.
Server location is usually treated as a latency question alone. If you serve an audience in Turkey, an Istanbul location saves 40–60 ms on every request compared with overseas — a difference that is not negotiable for game and voice servers.
But location has a second, less-discussed dimension: it is a legal decision. If you run a service processing personal data and your server sits abroad, that constitutes a cross-border transfer of personal data. Under Turkey's KVKK, three routes make it lawful: transfer to a country the Board has granted an adequacy decision, appropriate safeguards (standard contracts, binding corporate rules, or Board-approved undertakings), or narrow exceptions.
In practice the standard contract is the common route, and it carries an obligation that is easy to miss: a signed standard contract must be notified to the Board within 5 business days of signature. Failure to notify carries administrative fines. Transfers through cloud systems, mail servers and SaaS tools all fall within scope.
Hosting in Turkey removes that chain entirely. Separately, under Law No. 5651 hosting providers must retain traffic data for six months; working with a domestic provider makes that compliance — and producing records when requested — considerably more practical.
This section is general information. Consult a legal advisor about what your specific data processing activity requires.
This is the decision non-technical buyers skip most often. On an unmanaged VDS the provider hands you a working server and network; operating system updates, firewall, backups, panel installation and application troubleshooting are all yours.
Without system administration experience you have three options: install a control panel and learn as you go, buy system administration support on a retainer, or move to a product with no administration burden such as shared or reseller hosting. Comparing only the server price, without pricing those three, is misleading.
We covered the resource-allocation difference between VDS and VPS, and which workload needs which, in detail in our guide on the difference between VDS and VPS.
When evaluating a provider, get written answers to these:
Uptime percentages look similar in marketing copy but translate into very different amounts of downtime. Allowed monthly downtime:
Two things matter more than the percentage itself: what the commitment covers (network only, or hardware too; is planned maintenance excluded) and what happens when it is breached. An uptime figure with no remedy attached is just a sentence in a contract.
Server cost is not the monthly sticker price alone. When calculating total cost, add:
Watch for first-month discounts followed by much higher renewal pricing; compare annual totals instead. Annual billing usually earns a meaningful discount, but committing long-term before testing a provider is rarely wise.
The moment the server is handed over it is on the public internet, and automated scanners find it within minutes. Day-one tasks:
For a detailed checklist, follow our guide on the first 10 steps to secure a new VDS, and for the DNS side, our DNS records guide. If you are migrating, remember to lower your DNS TTL a day in advance.
Nubitro's VDS range runs on AMD Ryzen 9 9950X processors in Istanbul, Turkey. Because resources are assigned through hardware partitioning, CPU and RAM allocation is fixed per instance and does not fluctuate with neighbour activity.
Configurations start at 2 cores / 4 GB RAM / 40 GB NVMe M.2 SSD and scale to 12 cores / 64 GB RAM / 180 GB NVMe, with 6, 8, 10, 12, 16, 24 and 32 GB RAM tiers in between. Every plan includes 1 Gbps unmetered traffic, and the infrastructure carries 200 Tbps+ DDoS protection as standard. We explained what those DDoS capacity figures actually mean in practice in our DDoS protection capacity guide.
Review current configurations and pricing on our VDS/VPS server page, or move to dedicated servers for workloads that outgrow virtualization.
Yes — that is the right approach. Starting oversized simply wastes money. What matters is asking, before you buy, how the provider handles upgrades: if resources can be increased in place, growth is painless; if every upgrade means migrating to a new server, your planning changes.
4 GB for a small-to-medium site or development environment, 8 GB for mid-sized WordPress/WooCommerce or a small game server, 16 GB for busy e-commerce. If utilisation stays above 80%, move up a tier.
Three reasons: a 40–60 ms latency advantage for visitors in Turkey, no cross-border transfer obligations if you process personal data, and local support and billing. If your audience is entirely overseas and you process no personal data, an overseas location can be perfectly reasonable.
Run top on Linux and read the %st (steal time) value on the CPU line. If it stays above 10% during busy hours, the physical host is overloaded. With genuinely reserved resources it stays near zero.
Yes. Provider backups protect against infrastructure failure; your own copies protect against a database you deleted by mistake or an application bug. Never tie critical data to a single backup point.
When you consistently max out even the largest VDS plan, when you need GPUs or specific hardware, or when compliance requires single-tenant hardware.
Turkey (Türkçe)
Worldwide (English)
