Equipment / Dell Technologies / Storage
Each piece of data on the platform that suits it.
Volumes are increasing. However, the requirements vary from one dataset to another.A transactional database, user files, medical images, backups, and datasets for artificial intelligence do not all require the same level of performance, availability, or retention period. With Dell Technologies storage solutions, we design the architecture that best suits your applications and data.
Capacity isn't the most important issue
The choice of a storage platform doesn't depend solely on the number of terabytes. It depends on how data is created, accessed, modified, shared, protected, and retained —and these six verbs don't refer to the same machines.
A database responds; it does not store data.
A large number of short operations, each of which must complete in a few milliseconds. What matters is not the volume but the latency—and its consistency during peak hours.
We measure in milliseconds, not terabytes.
File sharing must remain accessible
Documents, shared folders, multimedia content, scientific data: many users and applications navigate the same directory structure, and it keeps growing without warning.
What matters is being able to grow without having to start all over again.
An artificial intelligence dataset contains millions of files
Millions of files or objects to browse quickly and store for a long time. The total throughput matters more than the latency of each individual access.
What matters is feeding data into the calculation without making it wait.
Three Ways to Present Data
Before we even get to the product names, there’s this question: in what format does the app want its data? The answer rules out two-thirds of the catalog, and it can be summed up in a single sentence.
Volumes that the server recognizes as disks
Storage capacity is presented as volumes that can be directly utilized by servers and their operating systems. Control over resources is precise, and performance is predictable.
For databases, virtualized environments, and transactional applications that require low latency.
A directory structure that everyone shares
The data is organized into folders and files, which can be accessed using standard network protocols. Multiple users and applications can work on the same structure at the same time.
For documents, shared folders, multimedia content, and scientific data.
Each data point has its own identifier and metadata
Each piece of data is stored along with its metadata and a unique identifier. Applications access it through a programming interface rather than by navigating a directory tree.
For modern applications, archives, very large volumes, and analytical datasets.
A single company rarely uses just one of these three forms. Blocks for databases and virtual machines, files for users, and objects for archives: this is the norm. The challenge is not to reduce the number of forms, but to identify what can be consolidated onto a common platform and what requires a specialized platform.
Six platforms, six goals
They are not ranked from worst to best. Each one serves a specific purpose, and the most expensive option is a poor choice if it doesn't meet that purpose.
Simple and Accessible Block Storage
Storage and direct-attach networking solutions designed to expand server capacity, host business applications, or support a mid-sized virtualization environment.
The choice whenease of deployment and cost control take precedence over everything else.
Unified All-Flash Storage
Block and file storage combined on a single, all-Flash platform using NVMe, with support for virtual volumes from VMware. Low latency, high availability, and a single console to learn.
The choice whenblock and file needs can be combined without compromising performance.
File storage that scales by nodes
A scalable network storage architecture designed for large volumes of unstructured data. Capacity and performance can be scaled by adding nodes, without disrupting the architecture.
The decision to make this choice whenthere are too many files for a single file server to handle.
Distributed object storage
A distributed object-oriented architecture for modern applications, data lakes, archives, and very large volumes of unstructured data. Access is provided through a programming interface compatible with the industry standard.
The choice when the applicationis object-oriented, or when storage spans years.
Software-Defined Block Storage
The resources of multiple nodes are combined into a distributed blockchain platform. Computing and storage can scale independently, which is not possible with an integrated architecture.
The challenge arises whenthe growth in computing power and storage capacity do not keep pace with one another.
Block Storage for Mission-Critical Environments
Dell Technologies 's high-end block platform: mission-critical applications, very large databases, financial systems, and mainframe environments requiring enhanced continuity.
The choice whenan outage is measured in consequences, not in minutes.
We help you identify the platform that best suits the nature of your data, your applications, your performance requirements, and your future plans. The goal is not to use a multitude of technologies, but to build a cohesive architecture where each piece of data is placed according to its value.
The installed space is not the space you will be able to use
Three mechanisms allow more data to be stored than the physical capacity would suggest. They work, but not for all types of data —and it is this second condition that determines the sizing.
Deduplication keeps only one copy
Identical blocks are recognized and stored only once. In a virtualized environment, where fifty machines share the same operating system, the benefit is considerable.
Compression reduces file size before writing
Data is reduced in size before it reaches physical storage media. The effect depends entirely on the nature of the data: text can be compressed, but a video that has already been compressed cannot be compressed further.
Space is allocated only when data arrives
The application receives the capacity it requests, but that capacity is allocated only as data is actually written. A committed reserve ceases to be a locked-in reserve.
It is also the mechanism that allows us to exceed our actual capacity without realizing it: it requires monitoring, not just activation.
That which cannot be reduced, or can be reduced only very slightly
Videos, pre-compressed images, encrypted data, and certain backup formats offer little to no potential for reduction. Virtualized environments and systems containing many similar files, on the other hand, yield significant reduction.
Therefore, capacity planning is never based on a single ratio. It takes into account the gross capacity, the remaining capacity after protection, the type of data, its growth, and the growth rate that can reasonably be expected.
| Platform | Announced rate | What this entails |
|---|---|---|
| 3rd-Generation PowerStore | 6:1 | Data Eligible Under the Dell Program |
| 2nd-Generation PowerStore | 5:1 | Data Eligible Under the Dell Program |
| PowerMax | 5:1 | Data Eligible Under the Dell Program |
| PowerFlex | 2:1 | Data Eligible Under the Dell Program |
| PowerScale | 2:1 | Data Eligible Under the Dell Program |
These rates are those of the Dell Future-Proof program, and each is subject to the eligibility requirements specific to its platform. To illustrate the first point: sixty terabytes of data might occupy only about ten terabytes of physical capacity, provided the data and terms of use meet the warranty requirements. These are program terms, and they are subject to change; we double-check them before each proposal.
Source: Dell Future-Proof Program
What is your situation?
Twelve starting points. They don’t lead to the same platform or the same budget —and identifying yours is the key to winning the first meeting.
Your capacity is running out
Volumes are growing faster than expected, and your current infrastructure no longer allows you to easily add capacity.
Your apps are running slowly
Databases, virtual machines, and business applications require more operations per second, lower latency, or higher throughput.
The number of your virtual machines is increasing
The platform must deliver consistent performance and high availability, and remain manageable even as the number doubles.
Your storage system is too old
It is nearing the end of its support period, or it no longer meets current performance, capacity, and security requirements.
Your data is scattered
They are spread across multiple systems, servers, and file-sharing solutions. You want to reduce the number of platforms and centralize their administration.
Your file sizes are growing very quickly
Documents, images, videos, scientific data, and business content require an architecture that can scale without adding another file server.
Your development projects require object storage
Modern applications, data lakes, archives, and datasets require a standard programming interface and a distributed architecture.
You are implementing artificial intelligence
Training, inference, and analysis require an infrastructure capable of powering computations without delay, while simultaneously improving capacity and throughput.
You operate multiple locations
Your data must be distributed, replicated, or accessible from multiple locations, while maintaining consistent management.
A system failure shouldn't cost you your data
Your mission-critical applications require replication, snapshots, and continuity mechanisms to minimize the impact of an outage.
You need to move your data
The project must take into account the volumes, dependencies between applications, available windows, and the acceptable downtime for each.
Your data storage costs are rising
Optimize physical capacity, data reduction, performance, and lifecycle—to avoid oversizing that ends up costing you for five years.
Moving data doesn't fix what was wrong
Replacing a storage platform is not simply a matter of transferring data from one system to another. A poorly planned migration replicates existing limitations on new hardware: poorly organized volumes, overlooked dependencies, unrealistic windows, and capacity calculated without accounting for growth.
Assess the existing environment
Applications, servers, virtual machines, protocols, volumes, performance, dependencies, and availability requirements.
This report specifies which data should be moved, which can be consolidated, which should be reorganized, and which should remain on a specialized platform.
Designing the Target Architecture
Storage cannot be considered in isolation: performance depends as much on the platform as it does on the servers, the network, the hypervisors, the applications, and the security mechanisms.
Each choice is linked to a measure from the previous step. Those that are not linked are excluded from the project.
Move in stages, with checkpoints
The pace is determined by the volumes to be transferred, the available bandwidth, maintenance windows, and the acceptable downtime for each application.
Each step concludes with a check: integrity, access, performance, and the application's behavior on the new platform.
Simplify day-to-day operations
Documentation, monitoring, alerts, replication, snapshots, backup, and disaster recovery procedures are part of day-to-day operations—not an appendix.
This migration is the only time when all of this can be set down neatly without interrupting anyone.
The scope is defined before we begin. This includes what we’ll handle, what you’ll retain, the option to reverse course, and a clearly stated limit. The same engineers will support you from the site survey through commissioning.
What Complements Your Storage
A storage platform never comes alone. Here's where to find the rest.
Servers that use this storage
The four product lines— PowerEdge —and the chain that connects what you measure to what you buy.
View servers Dell TechnologiesThe entire Dell product line, from computers to datacenter
Networking, hyperconvergence, data protection, and preparing hardware before shipment.
View the entire collection Data ProtectionWhat protects the data you store
Two time frames shape the entire architecture, and a single barrier alone answers the question of the day of the attack.
See Data ProtectionAn on-premises artificial intelligence project starts with storage. The corpus, index, and logs of a private RAG are stored on this storage system, along with the backup and rollback procedures described on the private artificial intelligence page.
Let's talk about your storage
Tell us about your data, and we'll tell you where it should go
What you're storing today, how your applications access it, and how long you need to keep it. With these three elements, the first meeting is already productive.
What we offer is a meeting with the engineers who will carry out the work. Volume surveys and sizing are part of the scope of work. The preliminary discussion, however, is free and is often enough to determine whether the project is feasible.
Renens, Sion, Châtel-Saint-Denis
Microsoft Gold Partner and a Gold Partner Dell Technologies. A single point of contact for sizing, ordering, migration and support.
Renens VD +41 21 806 37 15
Sion VS +41 27 552 00 22
Châtel-Saint-Denis FR +41 26 322 59 05