---
title: "HPE SimpliVity case study: from 20-hour backups to seconds"
author: "Abdulaziz Akyol"
author_url: https://www.abdulazizakyol.com/en/about/
url: https://www.abdulazizakyol.com/en/blog/hpe-simplivity-case-study-from-20-hour-backups-to-seconds/
language: en
published: 2021-04-05
updated: 2026-09-25
categories: ["Blog", "Systems administration", "Leadership"]
tags: ["HPE SimpliVity", "hyperconverged infrastructure", "HCI", "backup", "data deduplication", "retail IT", "case study"]
translation: https://www.abdulazizakyol.com/blog/hpe-simplivity-projemizin-basari-hikayesi/
description: "Moving Civil Mağazacılık to HPE SimpliVity: 2.9:1 deduplication, 1.5:1 compression, over 1 PB of backups and a backup window cut from 20 hours to seconds."
---

# HPE SimpliVity case study: from 20-hour backups to seconds

By [Abdulaziz Akyol](https://www.abdulazizakyol.com/en/about/) · 2021-04-05

## Key takeaways

- Instead of refreshing ageing server and storage hardware, Civil Mağazacılık moved its data centre to HPE SimpliVity hyperconverged infrastructure, with HPE Türkiye and partner Mayasoft.
- When all data from raw storage was moved to SimpliVity, the live system reached 2.9:1 deduplication plus 1.5:1 compression, with no performance loss.
- Backups that used to take up to 20 hours a day dropped to seconds; with more than 1 petabyte of versioned backups, capacity management ran at 98.5% efficiency.
- HPE's leasing (OPEX) model was chosen over purchase, and part of the existing estate stayed in use in a hybrid architecture.
- The real lessons: measure backups by recovery time rather than backup time, verify deduplication ratios on your own data and tie the infrastructure decision to a business goal.

This is a case study of the project that moved the data centre of Civil Mağazacılık, a Turkish retailer, to HPE SimpliVity hyperconverged infrastructure. I led the project as IT Manager, and HPE published it as a success story in Turkish in 2021. Every figure below comes from that document; the interpretation and the lessons are mine.

## The context: infrastructure for a growing retailer

Civil Mağazacılık was founded in 1996 as Civil Bebek ve Çocuk Dünyası, a baby and children's retailer. When HPE's document was written, the company sold through 70 stores in Türkiye, more than 10 stores abroad and its civilim.com online shop. Across 200,000 square metres of retail space and with more than 3,000 employees, it offered around 42,000 products from nearly 650 brands every year.

At that scale, infrastructure means keeping store tills, e-commerce, logistics, the supply chain and reporting up at the same time. The pandemic blurred the line between physical stores and digital channels almost overnight and raised the bar even further. As I put it back then: "Retail is going through a period in which it breathes through technology."

We had a clear choice: refresh ageing servers and storage piece by piece, or change the architecture. We chose the second.

## Why hyperconverged infrastructure?

A traditional three-tier data centre has servers, a separate storage array (SAN) and the storage network connecting them. Every tier has its own management console, its own support contract and its own bottleneck. Hyperconverged infrastructure (HCI) combines compute, storage and virtualisation on standard server nodes under one software layer.

Our project goals were:

- Best-in-class digital accessibility
- Uninterrupted and secure infrastructure
- Single-point management
- Flexible, high-performing and simple infrastructure
- Efficient and cost-effective infrastructure

The reasons we chose SimpliVity mapped directly onto that list: simple management, the promise of up to 90% capacity savings, built-in data protection and very fast restores. My summary at the time was that SimpliVity let us reach a more efficient infrastructure at a better cost "instead of pouring money into ageing equipment".

## The results in numbers

| Metric              | Result                                         |
| ------------------- | ---------------------------------------------- |
| Deduplication       | 2.9:1 when all data from raw storage was moved |
| Compression         | 1.5:1 on top of deduplication                  |
| Performance         | No performance loss during deduplication       |
| Backup time         | From up to 20 hours a day to seconds           |
| Backup volume       | More than 1 petabyte, versioned                |
| Capacity management | 98.5% efficiency                               |

The most valuable row, in my view, is the performance row. Every storage product lists deduplication on paper. Doing it on a live system, at peak hours, without users noticing, is a different matter. I called it "one of the best results a company can get on its live system" then, and I still think so.

## How deduplication and compression work

Deduplication splits the data written to storage into blocks and fingerprints each block. If a block with the same fingerprint has already been written, it is not written again; only a pointer is added. Most virtual servers carry the same operating-system files, so repetition in these environments is high.

Compression then shrinks the repetition inside the unique blocks. The two complement rather than replace each other, which is why our result was reported as two separate figures: 2.9:1 deduplication, then 1.5:1 compression on top.

One caveat: these ratios depend on the data. Database backups and VM images deduplicate well; already-compressed video, photos and encrypted files barely shrink. The ratio in a vendor brochure is not your ratio, so do not plan capacity without measuring.

## The backup window: from 20 hours to seconds

Before the project, backups took up to 20 hours every day. In practice part of the infrastructure was busy with backups almost all day, and any hiccup risked colliding with the next day's run.

SimpliVity changed that because data was already deduplicated and compressed as it was written. Taking a backup no longer meant reading all the data again and copying it elsewhere; it meant recording the metadata of the blocks at that moment. Whatever the size of the data, backups of terabytes could be taken in seconds.

The natural consequence was keeping backups more often and with more versions. The total versioned backup volume at Civil grew beyond 1 petabyte, and the 98.5% efficiency SimpliVity delivered turned capacity management into something that almost ran itself.

## Running the project: partner, budget and a hybrid transition

We ran the project with HPE Türkiye and its partner Mayasoft. During the analysis the Mayasoft team did not treat the work as a hardware refresh but designed a solution in proportion to the company's future plans, which is how existing resources we could keep using were brought into a hybrid architecture.

On the budget side we chose HPE's leasing (OPEX) model over purchase. Turning a large capital expense into an operating cost spread over years made a significant difference to our budget. Getting the products delivered quickly despite the pandemic also kept the project on schedule.

HPE's list of benefits ended with single-point management of the whole system, fast backup and deduplication, single-point security and an infrastructure ready for data analytics and AI. That last item was a goal at the time. Later the same year I founded CX Teknoloji and started working on [real-time video analytics on existing cameras](https://www.abdulazizakyol.com/en/blog/video-analytics-on-existing-cameras-from-rtsp-to-events/), and saw from the other side how decisive data-ready infrastructure is.

## Five lessons for IT leaders

**1. Measure today first.** "Backups take too long" is a complaint; "backups take 20 hours every day" is the starting point of a goal. The only reason we can tell this story with numbers is that we knew the baseline before we started.

**2. Test recovery, not backup.** A fast backup is a nice result, but the real question is how long the system takes to come back after a disaster and how much data is lost. Write down your recovery time objective (RTO) and recovery point objective (RPO) and rehearse restores regularly.

**3. Verify ratios on your own data.** Deduplication and compression ratios vary widely by data type. Plan capacity from a pilot measurement on your data, not from the vendor's example.

**4. Treat the financing model as part of the technical decision.** CAPEX or OPEX matters as much as which product you pick, and the right financing model makes the right architecture easier to approve internally.

**5. Tie infrastructure to a business goal.** Our goal was not "new storage" but infrastructure that seamlessly supports both the physical stores and the digital channel. The move to Azure and the SD-WAN rollout in the same years served the same goal.

## How would I make the same decision today?

Looking back from 2026, I would stress two things differently. First, ransomware: today the first question about a backup architecture is not speed but whether there is an immutable copy isolated from production. Second, the cloud: hyperconverged infrastructure and public cloud are no longer rivals but two parts of a hybrid architecture in which each workload runs where it is most efficient.

What does not change is the method: measure, tie it to a business goal, verify with your own data, then decide. I have collected this and similar projects, with awards and sources, on the [Achievements](https://www.abdulazizakyol.com/en/achievements/) page.

## Frequently asked questions

### What is hyperconverged infrastructure (HCI)?

Hyperconverged infrastructure combines compute, storage and virtualised networking on standard x86 server nodes under a single software layer. Instead of a separate storage array and SAN, it is managed from one interface and scaled by adding nodes.

### How did HPE SimpliVity cut backups to seconds?

Because SimpliVity deduplicates and compresses data as it is written, a backup is not a full re-copy of the data but a record of the metadata of the blocks at that moment. That is why a backup of terabytes of data can be taken in seconds instead of the hours a traditional backup job needs.

### What does a 2.9:1 deduplication ratio mean?

Data that logically occupies 2.9 units is stored in 1 physical unit, so roughly a third of the disk space is enough for the same data. The ratio depends on the data: virtual machine images and backups deduplicate well, already-compressed media hardly at all.

### What was Abdulaziz Akyol's role in the project?

Abdulaziz Akyol was IT Manager (Head of IT) at Civil Mağazacılık at the time and owned the infrastructure decision and the migration; he tells the story on the company's behalf in the success story HPE published.

## Sources

1. [HPE — Civil chose HPE SimpliVity for strong digital customer loyalty (PDF, Turkish)](https://www.abdulazizakyol.com/uploads/2021/04/a00112522tre.pdf)
2. [HPE — hpe.com version of the success story (archived copy, 2 April 2021)](https://web.archive.org/web/20210402133835/https://www.hpe.com/tr/tr/pdfViewer.html?docId=a00112522)
3. [Wikipedia — Hyper-converged infrastructure](https://en.wikipedia.org/wiki/Hyper-converged_infrastructure)
4. [Wikipedia — Data deduplication](https://en.wikipedia.org/wiki/Data_deduplication)

---

Abdulaziz Akyol is the founder of CX Teknoloji (AI, computer vision and IoT). Canonical version: https://www.abdulazizakyol.com/en/blog/hpe-simplivity-case-study-from-20-hour-backups-to-seconds/