Backing Up Mountains of Data to Disk

5 Year Total Cost of Ownership Case Study

The following are the results of an IT Brand Pulse total cost of ownership study. The report includes actual price quotes from SAN solutions from EMC, IBM, and NetApp, compared with five software-defined storage solutions from DataCore, Red Hat, Scality, SUSE and VMware. The study analyzes the cost of hardware, software and support for 250TB of storage, growing 25% per year, and used for backup to disk. Detailed cost data is in the analyst report and can be downloaded here.

A Catalyst for Change: The Explosive Growth of Unstructured Data

Data that is neatly organized in a database management system is commonly referred to as structured data, while data that is not managed with a database is often referred to as unstructured data. Data storage priorities are being changed by billions of new internet connected device such as sensors, watches, smartphones, and cameras, which are spewing trillions of unstructured data files.

Although the unstructured data is hardly ever accessed, if at all, an offline backup copy is not good enough. The data must be stored online because it is grist for the mill of big data analytics engines inside today’s business intelligence applications and tomorrow’s machine learning applications. The result is IT organizations are using backup software to ship mountains of cold data to online archives. The priority for application environments using unstructured data is to minimize storage costs while meeting the growth in requirements for capacity.

The New Business of Backup: Storing ALL Data for Analytics

As big data analytics become a common component of business intelligence applications, analytics vendors are asking customers to store ALL their data in order to get the best results from their analytics engines.

As a result, the business of “backup” software vendors has changed to “data management.” Products from vendors such as Commvault, Dell EMC, HPE, Veeam and Veritas are aiming to track all the data in an enterprise, the applications that are using it, where it came from, where it’s going, and who is accessing it.

The role of core backup applications has expanded to managing ALL the data in an enterprise, and the requirement for backup-to-disk is growing because real-time analytics needs the data to be online.

Looking forward, even small and medium will be trending towards storing ALL of their data online in order to leverage analytics capability that will exist in every business intelligence application.

Storing All Data with Software Defined Storage

All major data management and protection vendors have the ability to backup data from application servers to disk arrays. In case you were not aware, a server is a software or hardware device that can accept and respond to requests made over a network. The device that makes the request, and receives a response from the server, is called a client. On the Internet, a server also refers to a computer system that receives a request for a web document and sends the requested information to the client. As these SeedboxCo reviews explain, different servers have different levels of privacy and security protections, and therefore doing your research is crucial if you want to find the best server for your specific needs. That being said, servers have been in use for some time now. What’s relatively new though is the ability to backup data to software defined storage based on industry standard servers and object software defined storage. There has been a large increase in the amount of people using cloud server hosting to ensure their documents are protected and easily accessible in the cloud. The new generations of solutions are far more open and scalable than previous generations of backup-to-disk solutions based on SAN storage. This case study will reveal if backup-to-software defined storage is more cost-effective. Data-hungry businesses, in particular, may also wish to learn more about Colocation as an effective alternative.

The software defined solutions evaluated in this report incorporate an architecture which uses object storage and industry standard servers. The solution shown below uses SUSE Enterprise Storage OSD and monitor nodes running on HPE industry standard servers.

The Storage Systems

This report examines three disk array systems and four software-defined storage systems, including server hardware needed for hosting the software defined storage applications. The hardware, software and service products used in this case study all met the requirements specified in the RFQ. Differences in the products and services are described in the product overviews.

Entry-level disk arrays were utilized because they met the performance, availability and useable capacity requirements of the backup to disk application evaluated in this report. If performance-oriented mid-range or high-end storage arrays were used instead, the five-year TCO would have been significantly higher.

Dozens of features could have been added to all the configurations to enhance the performance (SSD), availability (replication) and usable capacity (compression and dedup). But a simple storage configuration met our report’s requirements for bulk storage, which is infrequently accessed.

5 Year TCO Numbers

Dell EMC Unity 300 – Five-Year Cost of Ownership: $330,865

The Unity 300 may be Dell EMC’s most affordable array, but that is like saying the C-Class is Mercedes’s most affordable car starting at $40,000. In addition, this system is not optimized for bulk storage. You can scale but with only fifteen 6TB drives per chassis, this results in the need for at least 1 new chassis every year.

Red Hat Ceph Storage – Five-Year Cost of Ownership: $328,847

The low-cost hardware foundation used for Red Hat Ceph Storage is exactly the same as the configuration used for other software-defined storage solutions in this analysis. The result of the capacity-based software pricing model is the second most expensive solution for our requirement.

VMware VSAN – Five-Year Cost of Ownership: $258,151

If support was included in the price of the software license-like it is with some other software-defined storage products-Virtual SAN would scale cost effectively. But it’s not; and the result is the cost of support exceeds the cost of the base product over a five year period.

DataCore SAN Symphony – Five-Year Cost of Ownership: $245,824

DataCore software is offers industry-leading performance for database/transaction processing workloads, and is priced accordingly. The storage capacity-based licenses add-up over the 5-year period to make SANsymphony the 4th most expensive solution we evaluated.

NetApp FAS2554 – Five-Year Cost of Ownership: $211,534

NetApp FAS2554 gets top marks for ease-of-use and for offering high-density configurations which should make scaling mass storage cost effective. But customers are asked to pay a hefty premium for the NetApp brand. For example, add-on 8TB drives for a NetApp FAS2554 can run well over $2,000 – over four times the cost of add-on drives for industry-standard servers supporting software-defined storage.

IBM v5010 – Five-Year Cost of Ownership: $195,458

If our application needed the additional features rolled into the v5010, this product might represent the best value for the money. But for our application-which requires the most capacity for the lowest cost-software-defined storage solutions are the ones to beat.

Scality RING – Five-Year Cost of Ownership: $193,384

The solution from Scality is the second, best-priced solution we examined. The difference between Scality, and price leader SUSE, was the cost of support. It’s also worth noting the cost of Scality software is entirely front loaded, while the cost of SUSE software is spread out evenly over the five-year period.

SUSE Enterprise Storage 4 – Five-Year Cost of Ownership: $149,408

SUSE offers several layers of cost savings to enterprise storage IT professionals. The foundation is industry standard hardware. Layered on top, is storage software with comparatively low annual license fees spread out over the life of your storage. Finally, support is included in the cost of the software license. For applications generating a lot of data, SUSE Enterprise Storage 4 scales capacity but not cost.


Discovery #1: Software-Defined Storage Eliminates Branded Storage Taxes

IT organizations have shown a strong preference for branded storage. Everyone knows they’re paying a tax for the EMC or NetApp logo, but because they deployed the “Mercedes of storage arrays”, they also figure they won’t get fired when something goes wrong. This branded storage tax is applied to every disk drive a customer purchases during the life of the system, and is as much as 4x the cost of HDDs used in industry-standard servers and software-defined storage systems.

Discovery #2: SUSE is the Price Leader for Disk Backup

The chart below shows a side-by-side comparison of eight disk backup solutions based on pricing provided by the manufacturers or their channel partners. Surprisingly, 2 of the 3 most expensive solutions were SDS products (not surprisingly from traditional enterprise vendors), while the clear price leader for this class of solution was SUSE Enterprise Storage 4.

Discovery #3: The Future of Disk Backup is Software-Defined Storage

The data in this report indicates that traditional enterprise storage is under tremendous price pressure from software-defined storage available from vendors like SUSE, especially for capacity-driven solutions such as disk backup.

This report also reveals that Red Hat and VMware are positioning the price of their software-defined storage offerings at parity with traditional (expensive) enterprise storage solutions.

The Bottom Line

IT organizations looking to lower the cost of disk backup should evaluate software-defined storage solutions. Based on easy-to-service x86 servers, the technology is proven by hyperscale public cloud providers and can be deployed in private clouds for 1/3 the cost of branded storage.