sdecoret - Fotolia
A lot of suppliers claim to be “market leading” when it comes to software-defined storage (SDS).
But to assess those claims, organisations need to understand the technology and what it can do for them.
What is software-defined storage?
In essence, SDS is storage management that uses software to control the provisioning of data storage independent of the underlying hardware model, type or age.
Thus, it has to be completely hardware-agnostic and run on any hardware platform. It must also be designed to scale up and avoid bottlenecks. There should also be a separation of data and control paths.
The technology enables an organisation’s IT administrators to configure, aggregate and allocate shared storage resources through a single dashboard.
The benefits of software-defined storage
According to ESG’s 2017 European storage trends survey, almost 60% of organisations have begun to deploy or are evaluating software-defined storage. Another 21% of organisations are interested in SDS in the long term. Those that entirely ruled out the idea of software-defined storage were in the minority.
There are many reasons for organisations to embrace SDS. Among them are more flexibility, faster and simpler scalability, and improved automation.
SDS systems cans be configured and deployed in virtually any way, with organisations able to specify the type of hardware, or run them on virtual, cloud or container platforms.
Read more about software-defined storage
- The evolution of software-defined storage, from simple software products to those that offer scale-out, hyper-converged, NVMe, public cloud and container functionality.
- The software-defined approach to storage is catching on. However, for now, enterprises prefer preconfigured SDS products bundled with hardware for easier deployment.
This means that older storage systems can be re-used while enabling centralised data storage management and the introduction of newer technologies such as data deduplication, compression and encryption. In a lot of cases, this can be less expensive than buying newer storage hardware that includes these technologies.
This extends the life of existing storage, evades hardware lock-in and lowers the cost of ownership. Organisations can choose industry-standard x86 server and disk hardware and configure storage service options for users. This offers organisations greater flexibility as the software is decoupled from the underlying hardware.
Why buy SAN or NAS instead?
SAN and NAS hardware can be a better choice when providing dedicated storage for a single purpose as it allows the storage and software to be more fully integrated and self-contained.
It can also allow organisations to take advantage of native data management software that comes with the hardware – such as compression, data duplication and encryption – with no additional costs. If each storage hardware device is managed through its own interface it also eliminates a single point of failure.
SDS vs SAN or NAS?
When it comes to choosing between SDS and SAN/NAS organisations have to decide on whether to get a best-of-breed product that will give them more value and potentially higher cost savings, or alternatively go for the convenience of an all-in-one solution.
The choice for one over the other all comes down to the maturity and size of the IT department. If your enterprise has a large IT team, it may well go for SDS as it probably has the skills and experience to support multiple suppliers.
Paul Mercina, Park Place Technologies
On the other hand, smaller companies may be better with a single supplier because they don’t have the resources to manage the complexity associated with multiple suppliers.
Lifecycle and scalability are significant issues with NAS and SAN. Some suppliers will expect customers to pay upfront fees for licences and hardware before they implement a traditional NAS or SAN architecture. The customer could also expect to then pay for support and maintenance over the next several years. After that, you can expect to then pay more money for upgraded hardware and software.
Software-defined storage avoids this as it is hardware agnostic and will run on any x86 server – so avoiding supplier lock-in. SDS also allows the use of existing hardware, which drastically reduces costs and enables organisations to base buying decisions for each upgrade on business requirements and availability of products.
However, some would argue that putting SDS on commodity servers can be more of a problem than expected, so here it can make sense to use equipment from leading storage manufacturers, with the price tags these suppliers advertise.
Paul Mercina, head of innovation at datacentre services firm Park Place Technologies, says that over the longer term it’s harder to find use cases where traditional storage trumps SDS.
“The ability to upgrade software and hardware separately, combine different storage devices and leverage commodity hardware ultimately make SDS the preferred solution for organisations undergoing a data volume explosion, which is already a near-universal phenomenon in many industries,” says Mercina.
“As does operating a multiprotocol storage environment, automating storage management functions, scaling out more efficiently without the traditional data migration headaches, and so on. It will rapidly become too difficult to keep pace using traditional SAN and NAS systems.”
Key providers: Software-defined storage
Speeds and feeds: Fibre Channel (up to 32 Gbps), iSCSI (up to 40 Gbps), and Fibre Channel over Ethernet (FCoE) via connection to FCoE switch
Drive type and capacities: Any internal drives, external drives, external disk arrays, JBODs, solid-state disks (SSD), flash memories, and intelligent storage systems supported on Windows Server 2008, 2008 R2, 2012 & 2012 R2 may be attached to DataCore nodes
Protocols supported: Standard IP network interfaces for inter-node communications, console access, and asynchronous remote replication between nodes. Standard IP LAN connections for file sharing (NAS)
Deployment size (min/max): 1PB (petabyte) per node; 64 nodes per group
Supplier: Dell EMC
Product: ECS EX300
Speeds and feeds: 10GbE front end, 10GbE back end
Drive type and capacities: 1TB, 2TB, 4TB, 8TB
Protocols supported: TCP/IP
Deployment size (min/max): Unstructured storage up to 1,536TB (terabytes) per rack
Product: ONTAP Select
Speeds and feeds: Minimum 2x 10Gb Ethernet ports (4-, 6-, 8-nodes)
Drive type and capacities: 8-60 drives
Deployment size (min/max): 4-, 6-, or 8-node cluster, Up to 400TB raw per node
Product: SUSE Enterprise Storage
Speeds and feeds: 10Gb Ethernet (two networks bonded to multiple switches)
Drive type and capacities: All types of disks in JBOD configurations, or local Raid. Disks should be exclusively used by SUSE Enterprise Storage
Protocols supported: iSCSI, NFS, CIFS/SMB, RBD (Block), RADOS (Object), CephFS (With multiple active MDS Servers), S3 & Swift
Deployment size (min/max): SUSE Enterprise Storage scales to exabytes of data and maximum number of nodes is only limited by the network topology
Product: Veritas Access 3340
Speeds and feeds: 14x 1Gb Ethernet, 4x 10Gb Ethernet
Drive type and capacities: N/A
Protocols supported: Amazon S3, CIFS, NFS, SMB for file and object access, and iSCSI
Deployment size (min/max): starts at 700TB of usable capacity and scales to 2.8PB
Speeds and feeds: N/A
Drive type and capacities: All-flash, hybrid, all-disk pools Raid 10, N+1, N+2, N+3
Protocols supported: File: NFS 3.1, NFS 4, SMB 2.1, SMB 3 Block: Fibre Channel, iSCSI
Deployment size (min/max): N/A