DataCore Parallel Server Rockets Past All Competitors, Setting the New World Record for Storage Performance
HPC Today  |  Wire  |  June 15, 2016
Note: the wired news below has been filtered but not edited by HPC Today.

Measured Results are Faster than the Previous Top Two Leaders Combined, yet Costs Only a Fraction of Their Price in Head-to-head Comparisons Validated by the Storage Performance Council

Following a scorching run of world records, DataCore Software today rocketed past the old guard of high-performance storage systems to achieve a remarkable 5.1 million (5,120,098.98) SPC-1 IOPS[1] on the industry’s most respected head-to-head comparison — the Storage Performance Council’s SPC-1 benchmark. This new result places DataCore number one on the SPC-1 list of Top Ten by Performance[2].
 
To put the accomplishment into perspective, the independently-audited SPC-1 Result for the DataCore Parallel Server software confirms the product as faster than the previous top two leaders combined. The Huawei OceanStor 18800 V3 held the previous top spot with 3,010,007.37 SPC-1 IOPS[3], followed by the Hitachi VSP G1000 (with Hitachi Accelerated Flash) with 2,004,941.89 SPC-1 IOPS[4].
 
DataCore’s record-smashing results were achieved on a pair of standard Lenovo servers connected by Fibre Channel to 12 external hosts, each generating enterprise application workloads to drive the required I/O traffic (see the DataCore Parallel Server Full Disclosure Report[1]). The 2-node configuration used DataCore Parallel Server, a software solution that harnesses multicore computing to perform I/O processing in parallel, thereby transforming standard computing platforms into parallel storage servers well suited to driving I/O for demanding applications on the storage network as well as in hyper-converged stand-alone systems. 
 
“While the world record in performance is a great accomplishment, for DataCore, it is just the start of what can be achieved with Parallel I/Otechnologies and multicore advances – it is boundless,” said George Teixeira, president of DataCore Software. “These results were based on just two small standard servers. Our software flexibility allows more cores and more server nodes to be added, but our real objective was not just to show top performance. Instead, we wanted to demonstrate that we are in a new era of I/O and storage architecture driven by parallel software capable of harnessing the processing advances presented by a transformed but still unmistakable expression of Moore’s Law.” 
 
Best Price-Performance for an External Storage System 
For this level of performance, customers typically pay upwards of $2 million for the storage system. For example, the total price for the Huawei OceanStor system was reported at $2,370,763.89[3], while the Hitachi VSP system was $2,003,803.84[4]. The DataCore solution comes in at only $506,525.24[1]. DataCore offered an unprecedented price-performance of $0.10 per SPC-1 IOPS[1], making it a 7X times better value in terms of price-performance.
 
Affordable price-performance and accompanying lower operating costs enable a much broader set of customers to tackle high-performance and I/O-intensive problems such as data analytics and large scale databases within their budget and space constraints.
 
Fastest Response for an External Storage System 
Average response time under heavy loads is another telling indicator, especially for customers eager to shorten the time it takes to turn big data into actionable intelligence for a competitive edge. DataCore Parallel Server software sustained an unheard-of average of 280 microseconds (0.28 milliseconds) response through the entire eight hour SPC-1 Sustainability Test Run[1] — showing none of the elongated delays seen on many competing products. The previously top-rated Huawei and Hitachi systems’ average response times were more than 3X slower (0.92 milliseconds[3] and 0.96 milliseconds[4], respectively). 
 
This average response time is all the more impressive for an external storage system accessed over a conventional Fibre Channel SAN (Storage Area Network), where other products have experienced considerably higher latencies. With partners QLogic, Brocade, and Lenovo, the DataCore configuration proved that contemporary Fibre Channel SANs are no barrier to blistering storage performance.
 
“Lenovo is excited to be part of this DataCore solution, which is poised to disrupt the storage marketplace by providing customers the top performance and best price performance in the industry,” stated David Lincoln, general manager of Lenovo’s Storage Business Unit. “DataCore’s industry-leading SPC-1 results on Lenovo System x demonstrate the high levels of performance, innovation and reliability that Lenovo is delivering to meet the demanding storage needs to our customers.”
 
The Parallel Era is Here: DataCore Parallel Server Harnesses the Power of Multicore Computing 
“DataCore’s new Parallel I/O technology, and its high performance should attract potential customers that are looking to accelerate applications without breaking the bank. Multicore processing has been around for over a decade – but for the most part, software development hasn’t fully evolved to take advantage of the parallel processing capabilities that these powerful, multicore x86 systems can provide,” said Steven Hill, senior analyst for storage technologies at 451 Research. “If these benchmark results translate into real-world performance, DataCore Parallel Server technology could prove to be disruptive to both the systems and storage markets.”
 
The record-setting numbers clearly demonstrate that DataCore Parallel Server offers a revolutionary new software architecture for I/O-intensive, latency-sensitive applications. Its Parallel I/O technology transforms standard computing platforms into powerful parallel storage servers. DataCore executes many independent ‘no stall’ I/O streams simultaneously across multiple CPU cores, dramatically reducing the time it takes to process I/O by taking full advantage of cost-effective but dense multicore servers. This technology removes the serialized I/O limitations and bottlenecks that restrict the number of workloads that can be consolidated on a server — internally or externally — and instead enables them to process far more work per server and significantly accelerate I/O-intensive applications.
 
Powered by DataCore Parallel I/O Technology
With uncontested leadership in all three key SPC-1 dimensions, including overall SPC-1 IOPS, average response time at 100% load and overall SPC-1 Price-Performance[5], one can see how far DataCore Parallel I/O Technology has taken the state of the art in storage over the last year.
 
For additional results highlighting a multi-node hyper-converged configuration, please see our recent announcement: DataCore Up-scales Its Record-Breaking Hyper-converged Performance with Multi-node Highly Available Server SAN.
 
SPC-1 Full Disclosure (FDR) Report
The rigorous SPC-1 performance testing is designed to demonstrate a system’s performance capabilities for business-critical enterprise-level workloads typically found in database and transaction processing environments. The audited configuration that was tested and priced includes DataCore Parallel Server software running on two standard Lenovo System x3650 M5 servers. For complete configuration, pricing and performance details please see the full audited SPC-1 report at: http://www.storageperformance.org/results/benchmark_results_spc1_active/#a00179
 
Source: DataCore
 
 

[1] DataCore Parallel Server (Dual Node, Fibre Channel SAN) (Current as of 6/15/2016)
[2] SPC-1 Results – Top Ten by Performance (Current as of 6/15/2016)
[3] Huawei OceanStor 18800 V3 (Current as of 6/15/2016)
[4] Hitachi Virtual Storage Platform G1000 (with Hitachi Accelerated Flash)(Current as of 6/15/2016)  
[5] SPC-1 Results – Top Ten by Price-Performance (Current as of 6/15/2016)

© HPC Today 2024 - All rights reserved.

Thank you for reading HPC Today.

Express poll

Do you use multi-screen
visualization technologies?

Industry news

Brands / Products index