It’s all about analytics! From data, to insight, and then action. Analytics, along with Internet of Things, represent the foundation of Digital Transformation that is changing businesses, industries, and lives globally.
Fundamental to this transformation is ‘integrated’, ‘software defined’ many things including compute, storage, network and management. Through innovations, Cisco’s Unified Computing System (UCS) integrates these elements in to a cohesive programmable infrastructure that can scale as workloads demand. Cisco’s UCS Integrated Infrastructure portfolio further optimizes the application architecture solutions based on your use cases, your applications, and technology needs.
Cisco UCS Integrated Infrastructure for Big Data introduced few years ago has become a popular choice in the enterprises across major industry segments including agriculture, education, entertainment, finance, healthcare, manufaturing, insurance, public sector, retail, and service provider environments.
These reference architectures are supported with leading software partners including Actan, Cloudera, DataStax, Elastic, Hortonworks, IBM (BigInsights), MapR, Oracle (No SQL Database), Pivotal (HD and GPDB), Platfora, SAS, SAP (HANA Vora), Splunk, and others.
Recently, the business value of our Cisco UCS Integrated Infrastructure for Big Data solutions were quantified by IDC through independent research, which showed the following benefits for an organization:
The new reference configurations and single SKU bundles are listed in Tables 1,2,3 and 4 below. These configurations can be deployed as-is, or referenced as templates for building custom configurations, and used to scale to thousands of servers on demand through the use of Cisco Nexus® 7000 and 9000 Series Switches.
Table 1: Performance Optimized: High performance Hadoop, Scale-out databases and analytics solutions
| Base Solution SKU | UCS-SL-CPA4-P1 | UCS-SL-CPA4-P2 | UCS-SL-CPA4-P3 |
| Connectivity | 2 Cisco UCS 6296UP | 2 Cisco UCS 6296UP | 2 Cisco UCS 6332 |
| Servers in Base Solution SKU | 16 | 16 | 16 |
| Server SKU (C240 M4) | UCS-SPBD-C240M4-P1 | UCS-SPBD-C240M4-P2 | UCS-SPBD-C240M4-P3 |
| Processor | 2 Intel Xeon processor E5-2680 v4 | 2 Intel Xeon processor E5-2680 v4 | 2 Intel Xeon processor E5-2680 v4 |
| Memory | 256GB DDR4 2400MHz | 256GB DDR4 2400MHz | 256GB DDR4 2400MHz |
| Boot | 2 240GB EV SSD | 2 240GB EV SSD | 2 240GB EV SSD |
| Data | 24 1.2TB 12G 10K SAS HDD | 24 1.8TB 12G 10K SAS HDD | 24 1.8TB 12G 10K SAS HDD |
| Capacity | 29TB | 29TB | 43TB |
| IO Bandwidth | 4.1GBytes/sec | 5.5GBytes/sec | 5.5GBytes/sec |
| Scalability (no additional switches) | 80 (20Gbps per server) | 80 (20Gbps per server) | 26 (80Gbps per server) |
| Scalability with ACI | 1000s | 1000s | 1000s |
Table 2: Capacity Optimized: Hadoop, Scale-out databases and analytics solutions
| Base Solution SKU | UCS-SL-CPA4-C1 | UCS-SL-CPA4-C2 |
| Connectivity | 2 Cisco UCS 6296UP | 2 Cisco UCS 6296UP |
| Servers in Base Solution SKU | 16 | 16 |
| Server SKU (C240 M4) | UCS-SPBD-C240M4-P-C1 | UCS-SPBD-C240M4-P-C2 |
| Processor | 2 Intel Xeon processor E5-2620 v4 | 2 Intel Xeon processor E5-2620 v4 |
| Memory | 128GB DDR4 2400MHz | 256GB DDR4 2400MHz |
| Boot | 2 240GB EV SSD | 2 240GB EV SSD |
| Data | 12 6TB 12G 10K SAS HDD | 12 8TB 12G 10K SAS HDD |
| Capacity | 72TB | 96TB |
| IO Bandwidth | 2.5GBytes/sec | 2.3GBytes/sec |
| Scalability (no additional switches) | 80 (20Gbps per server) | 80 (20Gbps per server) |
| Scalability with ACI | 1000s | 1000s |
Table 3: Starter Configuration for test and development, and smaller deployments
| Base Solution SKU | UCS-SL-CPA4-S |
| Connectivity | 2 Cisco UCS 6248 |
| Servers in Base Solution SKU | 8 |
| Server SKU (C220 M4) | UCS-SPBD-C220M4-S1 |
| Processor | 2 Intel Xeon processor E5-2620 v4 |
| Memory | 128GB DDR4 2400MHz |
| Disks | 8 1.2TB 12G 10K SAS HDD |
| Capacity | 10TB |
| IO Bandwidth | 1.4GBytes/sec |
| Scalability (no additional switches) | 40 (20Gbps per server) |
Table 4: High Performance Configuration with SSDs
| Base Solution SKU | UCS-SL-CPA4-H |
| Connectivity | 2 Cisco UCS 6332 |
| Servers in Base Solution SKU | 8 |
| Server SKU (C220 M4) | UCS-SPBD-C220M4-H1 |
| Processor | 2 Intel Xeon processor E5-2680 v4 |
| Memory | 256GB DDR4 2400MHz |
| Disks | 8 960GB EV SSD |
| Capacity | 7.5TB |
| IO Bandwidth | 4GBytes/sec |
| Scalability (no additional switches) | 26 (80Gbps per server) |
Additional Information
Solution Brief (SKUs)
Cisco Big Data Portal
Cisco UCS Big Data Design Zone
The Business Value of Cisco UCS Integrated Infrastructure for Big Data
This 4th generation of the UCS Integrated Infrastructure for Big Data is a significant improvement in performance. For example, the Intel Xeon E5-2600 v4 (Broadwell-EP) processor changes provide up to a 24% performance increase among key processor arithmetic performance metrics over the previous Intel Xeon E5-2600 v3 processors. In addition, this platform also uses the DDR4 2400MHz DIMM RAM modules offering lower latency and higher clock speed. This shortens the time to fetch more or new data from the CPU. The most recent Cisco TPCx-HS results at the 1TB and 10TB results demonstrate the significance of these changes. Congrats to the team that made these changes possible.
See Cisco’s broadwell benchmark records athttp://blogs.cisco.com/datacenter/cisco-ucs-claims-nine-world-record-benchmarks-with-the-intel-xeon-processor-e5-2600-v4-family-5
Thank you for the valuable information.
I want to know how to calculate IO Bandwidth.
Hi Kimihiro,
IO bandwidth is calculated through various synthetic workloads such as IOzone, FIO and also Hadoop benchmarks such as TestDFSIO, and others. We also used Industry Standard benchmarks such as TPCx-HS to identify performance and price-performance.
Thanks a lot !