A server for 1C: how to choose a configuration for the number of users
Servix technical team · Updated
In brief. A 1C server is sized by the number of users: up to 10, one Xeon E-2xxx, 32–64 GB and two SSDs in RAID 1 are enough; for 10–30, you need a Xeon Scalable with 8–16 cores, 64–128 GB and NVMe in RAID 10; for 30–100, the application server and the DBMS are separated, with 128–256 GB; from 100, two Xeons with 24–32 cores each and 256–512 GB on a DL380 Gen11 or R760. A frequency of 3.4–3.8 GHz matters more than the number of cores, and from 15–20 active users it is time to move to the client-server option.
The request “we need a server for 1C” says almost nothing about the configuration: “Accounting” for five workstations and an ERP for one hundred twenty users are different classes of hardware, differing in cost several times over. Mistakes are made in both directions: some buy a two-processor platform with forty cores for a file-mode database where half the resources will go unused, while others put seventy people on an office computer with a single SATA drive. Below is a breakdown of what really affects 1C speed, and approximate configurations for four ranges of user counts.
What 1C performance depends on
File mode and client-server mode
This is the first decision, and the entire configuration depends on it. In file mode, the database resides in a shared folder, and calculations are performed on client machines: the server effectively works as a file share, and the network load and number of locks grow faster than the number of users. The file database format has a limit on the size of an individual table (4 GB), and an improper client disconnect increases the risk of database corruption. The approximate limit of applicability is 5–10 concurrent users and a small database.
The client-server option separates the roles: the 1C:Enterprise application server (a server cluster, rphost worker processes) and the DBMS — MS SQL Server or PostgreSQL in a build adapted for 1C. This setup removes the table size limit and provides managed locks, correct transactions and proper backups without stopping work. Starting from roughly 15–20 active users, moving to the client-server option is justified almost always — even if formally the file database is still coping.
Processor: frequency matters more than core count
A key feature of 1C is that a significant part of the work runs in a single thread. Posting a document, recalculating totals, generating a complex report — these are sequential operations, and a particular user’s response time is determined not by the total number of cores in the system but by how quickly one core processes their request. Therefore, in typical 1C scenarios, a processor with a base frequency of about 3.4–3.8 GHz and a moderate number of cores provides a more comfortable experience than a “wide” many-core CPU at 2.0–2.3 GHz, even though the latter looks more impressive in core count and total computing power.
The number of cores, meanwhile, determines how many such single-threaded tasks the system can run in parallel, that is, scalability by number of users. The practical rule is simple: first choose a high-frequency model, then the number of cores required for the planned load. Look at frequency-optimized lines: Intel Xeon E-2xxx for smaller systems and Xeon Gold with a high base frequency for medium and large ones. You can choose a specific model in the server processorssection.
One more thing: in the BIOS and the operating system, set the maximum performance profile instead of power saving — power saving regularly turns out to be the cause of “floating” latency that people then look for in the configuration.
RAM
In the client-server setup, memory is shared between the DBMS and the application server. The DBMS tries to keep the “hot” part of the database in cache: if the volume of data that users actually work with fits in memory, the disk subsystem almost stops being a bottleneck. The application server uses memory for worker processes, session data and the metadata cache — with many users and heavy reports, this amounts to tens of gigabytes. Memory must be registered with error correction (ECC): for a system that holds the accounting of the entire company, this is not optional. Headroom in free DIMM slots is important: memory expansion is the cheapest upgrade.
Drive bays
The 1C load on drives consists mainly of small random reads and writes, plus intensive transaction log writes. Latency is what matters here, not sequential speed: an NVMe drive provides orders of magnitude lower latency than an HDD and noticeably lower latency than a SATA SSD. The database needs server-grade drives with power loss protection and high write endurance; consumer SSDs in this role wear out within months. The classic layout: a mirror (RAID 1) for the OS, RAID 10 for the database files, and separate volumes for the transaction log and tempdb. RAID 5 should not be used for intensive writes. You can choose drives in the SSD and HDD drivessection.
Configurations with SAS/SATA drives need a hardware RAID controller with cache and cache protection (a battery or flash module) — without it, the write cache is disabled and performance drops several times over. For NVMe, software mirroring or a separate boot drive controller is used.
Networking
For the client-server option, the link between the application server and the DBMS server is critical: if the roles are split across two machines, it should be at least 10 Gbps, preferably a dedicated pair of ports. For client connections with the thin client, gigabit to the workstation is enough, but in file mode the network becomes the main constraint, because it carries the data itself rather than query results. It is also worth providing a dedicated management port (iLO for HPE, iDRAC for Dell) — remote diagnostics and rebooting without a trip to the office save hours of downtime.
Configurations by number of users
The table below is a guideline for an initial calculation. The exact configuration depends on the application solution (ERP and Trade Management have higher requirements than “Accounting”), the database size, the depth of history, and the number of background jobs and exchanges. RAM and drive values are approximate, with headroom for 2–3 years of growth.
Users | Scenario | Processor | RAM | Drive bays | Suitable models |
|---|---|---|---|---|---|
up to 10 | “Accounting”, Small Business Management; a file database or a simple client-server setup on one server, often combined with file storage and a domain controller | 1× Intel Xeon E-2xxx, 4–8 cores, high base frequency | 32–64 GB ECC | 2× SATA/SAS SSDs in RAID 1 or 2× NVMe | HPE ProLiant ML30 Gen10, Dell PowerEdge T150/T360 |
10–30 | Client-server option: 1C server and DBMS on one machine, database from single-digit GB to several tens of GB | 1× Xeon Scalable, 8–16 cores, frequency-optimized SKU | 64–128 GB ECC | 2× SSDs in RAID 1 for the OS + 4× SSD/NVMe in RAID 10 for the database | HPE ProLiant DL320 Gen11, Dell PowerEdge R360/R660 |
30–100 | Separate roles: a dedicated application server and DBMS server, or two virtualization nodes with redundancy | 1–2× Xeon Scalable, 16–32 cores | 128–256 GB ECC | NVMe in RAID 10, separate volumes for the transaction log and tempdb | HPE ProLiant DL360 Gen11, Dell PowerEdge R660 |
100+ | ERP or Integrated Automation, 1C server cluster, fault-tolerant DBMS, external storage, standby node for failover | 2× Xeon Scalable, 24–32 cores and more | 256–512 GB and more | Local NVMe only, or a storage system with 25GbE or 32G FC connectivity | HPE ProLiant DL380 Gen11, Dell PowerEdge R760 |
The form factor is chosen based on the room: if there is no server room with a rack, a tower design is suitable for smaller configurations — it is quieter and better tolerates office conditions. If there is a rack, rack models are chosen: they are denser and easier to service. The full range of platforms is collected in the serverssection; the Dell PowerEdge line covers the same scenarios as HPE ProLiant — the choice is more often determined by the existing equipment fleet and preferences for the management system.
Typical mistakes
Most complaints about “slow 1C” come down not to the platform but to decisions made at the purchasing stage:
- Betting on core count instead of frequency. In 1C, a many-core low-frequency processor loses to a high-frequency one at the same price.
- Sticking with the file database “to the last.” 25–30 users in file mode means guaranteed locks and a risk of database corruption.
- Ordinary HDDs or desktop SSDs for the database. No power loss protection, low write endurance, unpredictable latency.
- RAID 5 for a database with active writes. The write penalty eats up the capacity gain; for databases, the right choice is RAID 10.
- Non-ECC memory and a desktop platform. Savings at the start turn into unexplained failures and no warranty for round-the-clock operation.
- One power supply and one drive. Any failure means accounting stops for as long as it takes to purchase a replacement.
- Virtualization without resource reservation. A 1C server on an oversubscribed host next to “noisy neighbors” runs unstably.
- No scheduled DBMS maintenance. Without statistics updates, reindexing and transaction log management, the database degrades within a few months.
- No backups with restore testing. A backup that has never been restored is not a backup.
What to consider in Uzbekistan
Power. Voltage sags and brief outages are a reality in many offices and industrial zones. At a minimum: a server with two power supplies connected to different lines, a UPS with enough runtime for the DBMS to shut down properly (approximately 15–30 minutes), and automatic shutdown configured on a UPS signal. If downtime is unacceptable — a voltage stabilizer at the input and a generator.
Cooling. In summer, the outdoor temperature in Tashkent stays around +40 °C, and the server room becomes a critical space. Rack servers are designed for an inlet air temperature of approximately 10–35 °C (ASHRAE class A2), with a comfortable range of 18–27 °C. When overheating, the fans go to maximum, the processors lower their frequencies, and “1C slowdowns” appear precisely in the heat. In practice: a dedicated room, two alternating air conditioners instead of one, temperature sensors with alerts, dust control and regular cleaning of filters and heatsinks.
Delivery, warranty and payment. Configurations are made to order with a lead time of 2–3 weeks, and the price is fixed in the invoice, so the timeframe needs to be built into the project in advance. What matters is original components with a warranty of 12 months, local service support and the ability to get replacement components quickly. For companies, we work by bank transfer with VAT and provide a contract, a VAT invoice and a full set of documents for accounting.
Related collections and tools: Server for 1C.
Questions and answers
Can a regular computer be used instead of a server if there are only five users?
Technically, a file database will run on a powerful PC. But a desktop has no error-correcting memory, no redundant power, no hardware RAID and no remote management, and its warranty is not designed for round-the-clock operation. For five users, a sensible compromise is a small entry-level tower server: it costs not much more than a good PC but provides ECC memory, a drive mirror and service support.
MS SQL Server or PostgreSQL — which to choose for 1C?
Both DBMSs are officially supported by the platform. PostgreSQL in builds adapted for 1C removes the question of DBMS licensing costs but requires more careful tuning and expertise for ongoing maintenance. MS SQL Server is easier to operate and maintain, but its licenses noticeably increase the project budget. The choice has a moderate effect on hardware requirements: both DBMSs need fast drives, plenty of memory and a high processor frequency.
Is one powerful server better than two smaller ones?
Up to about 30 users, one server is economically justified; beyond that, it is worth looking at two. Separating roles (application server and DBMS) removes competition for memory and drives, and two virtualization nodes provide fault tolerance: if one of them fails, accounting keeps running, and scheduled maintenance takes place without stopping the company.
If you need a specific configuration for your application, database size and number of users, we will calculate several options with justification and prices, including UPS sizing and room requirements. Call: +998 92 182 1144, write to [email protected] or message us on Telegram at the same number — we will select a server for the task, not for the price list.
Related reading
Database server for PostgreSQL, MS SQL and 1C: how to choose and how much it costs
Server for a DBMS: MS SQL per-core licenses, memory for the database size, SSDs in RAID 10, the log on a separate volume, battery-backed cache. Builds and prices as of November 2026.
Server RAM: RDIMM, channels, ranks, DDR4 and DDR5
How many memory modules to install in a server, why it is important to populate all channels, how RDIMM differs from LRDIMM and MRDIMM, DDR4 or DDR5, how many GB to take for the task.
Server drives: HDD or SSD, SAS, SATA or NVMe — what to buy for the DL380 and R750
Which drives to buy for an HPE DL380 or Dell R750 server: SAS, SATA, NVMe, PERC and Smart Array RAID controllers. Prices in UZS, three ready-made sets. Tashkent.
How to choose a server: 1C, virtualization, file server, database
A practical guide to choosing a server: what matters for 1C, a virtualization host, a file server and a database. Common mistakes when selecting a configuration.




