Is Dual-Socket Server Performance Possible in a Single-Socket Server? Gartner Says Yes.

Imagine you are a small cafe owner who brews 20 gallons of tea every day for your customers. Your cafe has recently seen an influx of business, so you are going to invest in some new pots. Would you choose to buy a 100-gallon pot for your 20-gallon tea needs? Probably not. You would most likely purchase a 30 or 50-gallon pot that would not only help you grow, but would also be cheaper.

The same goes for IT. In many cases, businesses often purchase dual-socket servers without thinking about the workload requirement. Why should you spend more money for capacity that will be underutilized?

According to a Gartner analysis, “By 2021, dedicated x86 single-socket servers will be able to address 80% of the workloads in use in enterprise data centers, up from 20% in 2018”. The common cause is the limitations on memory and input/output (I/O) capacity of individual CPUs. To ensure applications run effectively, businesses are forced to over purchase processing capacity. On older CPU generations, it often requires two CPUs to meet the core counts and the memory requirements.

With advances in processor technology, a single-socket server can achieve the right performance for the right price. This is one of the reasons why we believe Gartner recommends using a dedicated single-socket server for specific workloads.

The EPYC™ CPU from AMD, for example, delivers greatly increased CPU core count, large memory capacity, and 128 PCIe lanes for bandwidth. Prior to the introduction of this architecture, users would have had to purchase two CPUs to gain these capabilities.

Budget is always limited, but don’t compromise performance.

The Dell PowerEdge R6415 is designed with customers in mind for two simple reasons.

  1. Boost CPU Utilization with Lower Total Cost of Ownership (TCO).

The Dell PowerEdge R6415 is equipped with one AMD EPYC processor with up to 32 cores. It can easily process more workloads faster than your old dual-socket system.

Besides saving on the CPU cost, you will also accrue savings from reductions in heat sinks, fans and other related commodities.

The potential savings on software license and subscription fees can be substantial. Take VMware vSan as an example, the licensing cost of a single-socket server compared to a dual socket server can save you up to $11,760 per socket, per node.[1]

  1. Faster memory and increased capacity.

The Dell PowerEdge R6415 can support up to 16 DIMMs and up to 10 NVMe drives. It reduces latency and helps fully utilize the high CPU core counts. This enables your organization to process more workloads faster.

Conclusion

The new Dell PowerEdge single-socket servers aim to provide a cost-effective solution for users to simplify their IT, improve productivity and increase efficiency.

Dell has a full portfolio of single-socket servers to meet your needs. Visit PowerEdge Servers on the Dell website to learn more about the full portfolio, including Intel-based single-socket servers and their different form factors (in both tower, and rack form factors).

Contact a Dell business advisor today to find out how our single-socket platforms can help you boost your IT efficiency.

[1] Tony Harvey, “Use Single-Socket Servers to Reduce Costs in the Data Center”, December 5th

About the Author: Minyi Pan

Minyi currently serves as a product marketing manager at Dell Technologies. Before entering the marketing field, she was a supply chain professional where her world revolved around analyzing data to optimize inventory, cost, and things arriving on time. Now she focuses on using what she learned in the supply chain field (like fancy data analytics) to share the best technology solutions and trends.