Storage Area Networks (SANs) can utilize both hard disk drives (HDDs) and solid-state drives (SSDs) for storage. The choice between HDDs and SSDs depends on the performance, capacity, and budgetary requirements of the organization.
What is a SAN?
A SAN is a dedicated high-speed network that provides access to consolidated, block-level storage. SANs are primarily used to enhance storage devices’ capabilities, such as availability, security, and flexibility.
SANs allow an organization to consolidate storage resources in data centers while still providing networks with high-performance storage access. SANs also offer advantages like storage virtualization, replication, and backup.
What is an SSD?
An SSD, or solid-state drive, is a type of high-performance storage device that uses integrated circuit assemblies and flash memory to store data persistently. Unlike a hard disk drive (HDD), an SSD has no moving mechanical components and provides much faster read/write speeds.
SSDs are commonly used in situations where high data access performance, durability, power efficiency, and reliability are critical. However, SSDs tend to be more expensive than HDDs for the same data storage capacity.
Do SANs use SSDs?
Yes, SANs can utilize SSDs as one of their storage mediums. The ultra-fast performance of SSDs makes them well-suited for SAN environments where low latency and high I/O speeds are required.
Here are some of the key reasons why SSDs are used in SAN architectures:
- Faster response times: SSDs have much lower read/write latency compared to HDDs. This accelerates applications and workloads running on the SAN.
- Higher IOPS: SSDs can handle a greater number of input/output operations per second (IOPS). This increases overall throughput and performance of the SAN.
- Improved scalability: The performance of a SAN does not degrade as much when adding more SSDs compared to HDDs.
- Lower power consumption: SSDs are more power efficient than HDDs, leading to lower electricity costs.
- Increased resiliency: SSDs have no moving parts, making them less prone to physical failure over time.
What are the benefits of using SSDs in SANs?
Using SSDs in SAN architectures provides several key advantages:
- Faster access to data – SSDs offer microseconds of latency, enabling SANs to provide high performance for transactional applications and workloads.
- Improved IOPS – SSDs can sustain higher input/output operations per second than HDDs, facilitating faster parallel access.
- Increased throughput – The sequential read/write speed of SSDs is much higher compared to HDDs.
- Enhanced scalability – SSDs allow SANs to scale up easier without performance degradation.
- Better utilization – SSDs reduce latency and improve IOPS, allowing SAN resources to be used more efficiently.
- Higher consolidation ratios – More virtual machines and data can be consolidated onto SSD-based SANs.
Overall, utilizing SSDs enables SANs to provide accelerated performance, support more workloads, and deliver faster access to data for applications.
What are some potential drawbacks of using SSDs in SANs?
Some potential downsides of using SSDs in SAN environments include:
- Higher cost – SSDs are generally more expensive per gigabyte compared to HDDs. Large SAN deployments using only SSDs can become costly.
- Lower capacity – HDDs are available in much larger capacities compared to SSDs. For bulk storage, HDDs may be more viable.
- Lifespan concerns – SSDs can wear out after a certain amount of writes. Careful monitoring and replacement is required.
- RAID overhead – RAID for data protection consumes some of SSD’s capacity and affects write performance.
- Encryption overhead – Full disk encryption supported on some SSDs also uses processing resources.
While SSD performance is superior, the considerations above should be evaluated when designing SSD-based SAN solutions.
Should SSDs be used for all SAN storage?
Using all-SSD storage for an entire SAN is usually not necessary or cost-effective. A hybrid approach combining SSDs and HDDs is generally better suited for most organizations’ needs and budgets.
SSDs should be deployed selectively for SAN workloads requiring the highest performance – such as databases, virtual desktop infrastructure (VDI), and mission critical applications. Bulk file storage, backups, archives, and less frequently accessed data can reside on lower cost HDDs.
A tiered storage architecture with SSDs in the performance tier and HDDs in the capacity tier provides optimal balance across cost, capacity, and speed.
All-flash SANs may make sense for organizations with specialized performance needs and sufficient budgets. But for most, hybrid SANs allow flash benefits to be realized where they matter most while HDDs store the majority of data in a cost-effective manner.
Guidelines for Effective Use of SSDs in SANs
- Use SSDs for frequently accessed data requiring low latency like databases.
- Utilize HDDs for infrequently accessed data, backups, archives, and file shares.
- Tier storage with SSDs for performance tier and HDDs for capacity tier.
- Evaluate RAID implementation to balance protection and SSD performance.
- Consider smaller capacity but higher performance SSDs for critical workloads.
- Plan for eventual SSD replacement/upgrade as utilization increases over time.
- Monitor disk utilization and prepare to scale SSD capacity as needed.
What are some examples of SAN vendors utilizing SSDs?
Many major SAN vendors offer all-flash or hybrid flash SAN solutions using SSDs:
Dell EMC
- PowerMax NVMe SSD array
- Unity hybrid flash storage
- XtremIO all-flash arrays
NetApp
- AFF A-Series all-flash SAN
- FAS hybrid flash SAN
- EF-Series all-flash arrays
Pure Storage
- FlashArray//X NVMe SSD SAN
- FlashArray//C mainstream SAN
- FlashBlade high-performance unstructured data storage
HPE
- 3PAR all-flash arrays
- Nimble Storage predictive flash arrays
- Primera NVMe SAN platform
Many options exist for all-flash and hybrid flash SAN storage from leading vendors, allowing optimal selection based on performance, capacity, and budget needs.
What network protocols and connections are used in SSD SANs?
SSD-based SANs utilize high-speed, low-latency networks and protocols to connect servers to the consolidated storage environment:
- Fibre Channel – Fibre Channel SAN is the most common type of dedicated storage network used with SSD SANs today. Provides lossless transport and speeds up to 16/32/128 Gbps.
- iSCSI – Runs SCSI storage protocol over cost-effective Ethernet networks. Used for more affordable SSD SAN configurations.
- NVMe-oF – Emerging standard implementing NVMe storage protocol over fabrics like Ethernet, FC, and InfiniBand for ultra-low latency.
- InfiniBand – High-speed low-latency network interconnect supporting SSD SAN connectivity with bandwidth over 100 Gbps.
- Ethernet – SSD SANs can utilize 10/25/40/50/100GbE in addition to Fibre Channel for SAN network links.
These high-performance networks provide the throughput, low latency, and lossless behavior required to get the most from ultra-fast SSD arrays in SAN environments.
How are SSDs physically installed in a SAN?
SSDs are packaged and installed into SAN storage systems in a few different ways:
- Drive sleds – SSDs are inserted into drive sleds and slid into empty bays in the SAN enclosure.
- Add-in cards – Some SAN systems use PCIe add-in cards that SSDs slot directly into.
- 2.5″ drive form factor – Smaller SSDs can be mounted into 2.5″ drive bays on the SAN chassis.
- M.2 form factor – Newer SSD gumstick modules connect directly onto motherboards or expansion slots.
For easy installation and replacement, most enterprise SANs use hot-swappable drive sleds. SSDs are typically protected by the SAN’s RAID data protection schemes once inserted.
Sample SSD SAN Physical Architecture
Here is an example architecture illustrating how SSDs can be deployed within a SAN array:
| Component | Description |
|---|---|
| SAN chassis | Houses SSDs, controller modules, fans, and power supplies in a rack-mountable enclosure. |
| SSD drive sleds | Enable easy installation of SSDs into empty bays in the chassis mid-plane. |
| Controller modules | Provide RAID, caching, processing, and SAN network connectivity. |
| Internal SSDs | High-performance SSDs provide primary storage capacity and throughput. |
| SAN networks | Connect servers to SSD-based storage over Fibre Channel, iSCSI, InfiniBand, Ethernet, etc. |
This demonstrates how SSDs can be packaged into enterprise SAN arrays to provide consolidated, high-speed storage delivered over dedicated SAN networks.
How is SSD performance optimized in SAN environments?
There are several best practices to optimize and maximize SSD performance in SAN implementations:
- Use enterprise SSDs with high endurance ratings and performance SLA guarantees.
- Ensure SSDs support SAN capabilities like dual-port connections and T10 end-to-end protection.
- Spread workloads across large SSD pools to increase parallelism.
- Enable SSD caching to boost performance of HDD pools.
- Tune RAID to balance protection and maximize SSD throughput.
- Isolate workloads by performance tier – SSDs for critical applications, HDDs for bulk storage.
- Monitor disk utilization and plan capacity growth ahead of time.
- Consider NVMe-oF to reduce latency compared to SCSI protocols.
Properly architecting, monitoring, and maintaining an SSD-enabled SAN improves availability and consistent performance for applications.
SSD SAN Performance Considerations
| Area | Factors |
|---|---|
| SSD selection | Endurance, latency, IOPS, throughput, dual-port, T10 PI support |
| RAID configuration | RAID level, stripe size, read caching settings |
| Pooling and tiering | Spread workloads, use of tiers, SSD caching |
| SAN network | Bandwidth, low latency, lossless behavior |
| Protocols | NVMe-oF, Fibre Channel, iSCSI |
| Monitoring | IOPS, throughput, latency, utilization |
Paying attention to these aspects will enable optimal SSD SAN performance for critical workloads and applications.
How does SSD caching in SANs work?
Many SANs use a tiered storage approach with SSD caching to optimize performance and costs. This works as follows:
- SSDs act as an upper performance tier, while HDDs provide a lower capacity tier.
- Frequently accessed “hot” data is cached on the SSDs.
- Infrequently accessed “cold” data resides on the HDDs.
- The SAN cache management software automatically controls tiering.
When a read request comes in:
- The cache is checked for the data first.
- If it is present on the SSD (cache hit), it is directly read from there.
- If not found (cache miss), data is retrieved from the HDD tier instead.
For writes, new data can be written to the SSD cache first. Data may later be destaged in the background from SSD to the HDD capacity tier.
SSD caching provides performance close to all-SSD SANs by optimizing the use of faster media for frequently accessed data while bulk resides on more economical HDDs.
Benefits of SSD Caching
- Faster perceived performance from SSD cache hits
- Cost savings compared to all-SSD storage
- Improved IOPS compared to only HDDs
- Background optimization of hot data placement
- Scaling capacity via cheaper HDD additions
With proper working set sizing, SSD caching delivers great performance for consolidated SAN workloads at reasonable cost.
Conclusion
SSDs provide transformational performance benefits for SAN storage environments. With dramatically faster speed compared to HDDs, SSDs accelerate applications and enhance overall workload efficiency on SAN infrastructures.
All-flash SANs provide consistently high performance but at high cost. Hybrid SANs with automated SSD caching tiers combine the speed of flash with the economy of HDD capacity. This balance makes SSDs accessible for broader SAN adoption while still delivering vastly improved performance.
Careful SSD selection, integration, and monitoring helps overcome drawbacks like cost or lifespan concerns. For workloads requiring the utmost speed, NVMe-oF further reduces latency compared to SCSI protocols on SSD SANs.
With their many advantages and rapid innovation pace, SSDs will continue growing as a pivotal acceleration tier for SAN environments well into the future.
