White Papers

Modeling Analytics for Computational Storage

Next generation flash storage will be armed with a substantial amount of computing power.

DOWNLOAD

Samsung Z-SSD and ScyllaDB: Delivering Low Latency and Multi-Terabyte Capacity in a Persistent Database

A comparative study of Scylla performance on Samsung Z-SSD and DRAM.

DOWNLOAD

A Real-time, Low Latency, Key-Value Solution Combining Samsung Z-SSD™ and Levyx's Helium™ Data Store

In this whitepaper, we present a high performance key-value store solution that can be used in a wide variety of latency sensitive applications tied to the rapidly expanding era of hyper-speed “big data.” By combining Levyx's high-performance Helium™ key-value store with Samsung's ultra-low latency Z-SSD™, we have demonstrated performance improvements of up to 10X over the conventional approaches of processing large-scale datasets.

DOWNLOAD

Scaling In-Memory Data Processing with Samsung Advanced DRAM and NAND/SSD Solutions

Samsung and VoltDB analyze advances in memory and in-memory database technologies to address challenges from massive increases in contextual data.

DOWNLOAD

High-Performance Workloads With Software-Defined Storage and NVMe SSDs

Over the past several years, cloud computing has proven to be a transformational technology, driving many improvements in how IT organizations deliver applications and services to their respective businesses. One major result of this shift to the cloud has been broader adoption of distributed computing approaches across an increasingly large portion of the enterprise portfolio.

DOWNLOAD

Enhancing the Value of Alluxio With Samsung NVMe SSDs

Alluxio, formerly Tachyon, is the world's first system which unifies data at memory speeds while achieving affordability through Alluxios innovative tiered storage functionality. This paper shows how Alluxio’s storage can be used with different storage media available in systems including NVME SSDs while providing in-line performance consistent with the speed of the underlying storage media. Alluxio provides the capability to leverage all the storage that is available in a system.

DOWNLOAD

ScyllaDB and Samsung NVMe SSDs Accelerate NoSQL Database Performance

This white paper evaluates the performance of ScyllaDB, an open-source, high-performance NoSQL data store, and Cassandra, a well-known NoSQL data store implemented in Java, when serving requests from memory as well as from a first-in-class Samsung NVMe SSD. To this end, we characterize system performance, as well as overall throughput and latency, using different workloads in the Yahoo! Cloud Serving Benchmark (YCSB).

DOWNLOAD

Performance and Endurance Enhancements with Multi-stream SSDs on Apache Cassandra

This white paper presents a highly efficient use case of Samsung multistream-enabled 12G SAS SSDs functioning within Apache Cassandra, and delineates comparative performance results for different workloads.

When using Samsung multistream-enabled 12G SAS SSD as backend storage for Cassandra, the primary benchmark findings are:

Cassandra write performance improves up to 300%
Cassandra average write latency decreases up to 67%
SSD WA factor decreases up to 66%, enabling the SSD to last three times longer

DOWNLOAD

Best Practices for MySQL with SSD

This paper underscores that all SSDs are not alike and that specific configurations and settings need to be considered to maximize performance and system value. In the software domain, MySQL Server is a commonly used Relational Database Management System (RDBMS). This whitepaper clarifies how today's data centers combine SSDs and MySQL to achieve a substantial business advantage. The paper focuses on Percona Server – a free, fully compatible, open source MySQL Server enhancement.

DOWNLOAD

Scaling Cassandra Performance with Datera Elastic Data Fabric and Samsung SSDs

This whitepaper examines the usefulness of Apache Cassandra 3.7 deployed on Datera Elastic Data Fabric. Key areas of system improvement included are:

Operational benefits with Datera EDF when deploying Apache Cassandra
Performance comparison of deploying Apache Cassandra on Datera EDF versus deploying Apache Cassandra on direct-attached SSDs (server local)
Ability to scale Apache Cassandra instances on Datera EDF

DOWNLOAD

Accelerating OLTP performance with NVMe SSDs

Today's datacenter transition to high performance, open-source and hyper scale architectures continues its relentless acceleration. This white paper contends that it is only natural for today's datacenters to combine NVMe SSDs and MySQL to achieve an unprecedented business advantage. Going one step further, we find Percona Server, a free, fully compatible, open source MySQL Server enhancement. Because Percona Server is especially optimized for the I/O subsystem, we were able to extract more throughput out of it when using fast storage devices, and therefore use it for the final experiments reported here.

DOWNLOAD

Performance Benefits of Running RocksDB on SSDs

There is an explosion of unstructured data hitting data centers and cloud storage servers across the world. Hard disk drives cannot meet the necessary performance requirements for big data while many SSDs are constrained by the throughput limitations of a SATA interface. Recognizing the difficulty, the industry is progressing towards a more scalable enterprise interface called NVMe. This unique whitepaper compares real world applications for workloads running on NVMe SSDs versus SATA SSDs using a persistent key-value store RocksDB to provide OEMs with a comparative benchmark performance analysis.

DOWNLOAD

Data Center PCIe SSDs: Smart Choice Results in 5x Difference in Performance

Solid State Drives have become core building blocks for servers and data centers. Choosing the right SSD can result in significant performance improvement and lower total cost of ownership. Tom's Hardware evaluates and compares PCIe SSD offerings from three leading vendors to help users make the best choice for their deployment.

DOWNLOAD

Dell PowerEdge R930 with Oracle: The Benefits of Upgrading to Samsung NMVe PCIe Storage

Responsive and high-performing Oracle database applications, required for smooth business operations, demand high I/O throughput from storage. This means that companies who run these applications need power servers with a range of options for upgrading storage. Additionally, the balance of storage and compute performance determines database system performance. Knowing that storage speed and technology is key, Dell designed the new four-socket PowerEdge R930 server with extremely fast storage technology in mind to deliver outstanding performance.

DOWNLOAD

NVMe SSDs and RDMA Deliver Hyperscale Performance – Windows Server 2012 SMB Direct Benchmark

Today's data center administrators contend with difficult and continually shifting cost/performance tradeoffs. Exploiting one performance accelerant – Remote DMA (RDMA) – remained operationally elusive because of its inherent configuration and operational complexity. That day is past. Today, Windows Server 2012 R2's award-winning SMB Direct feature provides native RDMA support.

Combined with Samsung's XS1715 SSDs and QLogic^® FastLinQ™ 45000 Series 100GbE RDMA adapters, scale-out server configurations accessing networked storage systems can now exhibit unprecedented performance:

More than 1 Million Storage Transactions per second
More than 19 GB/s of data movement capabilities

This whitepaper provides a brief benchmark methodology overview of such a system.

DOWNLOAD

NVMe SSDs and RDMA Deliver Hyperscale Performance – Linux iSEr Benchmark

Today's data center administrators contend with difficult and continually shifting cost/performance tradeoffs. Exploiting one performance accelerant – Remote DMA (RDMA) – remained operationally elusive because of its inherent configuration and operational complexity. Today, the Linux software stack provides native RDMA and iSER support by utilizing SCSI targets.

More than 3 Million Storage Transactions per second
More than 20 GB/s of data movement capabilities

Samsung NVMe devices provide groundbreaking levels of raw storage performance. With the advent of 100GbE RDMA adapters, systems are now able to harness this performance for real-world applications. This whitepaper provides a brief benchmark methodology overview of such a system.

DOWNLOAD

Peer-Reviewed Publications

2017

NVMe-over-Fabrics Performance Characterization and the Path to Low-Overhead Flash Disaggregation

Abstract:
Storage disaggregation separates compute and storage to different nodes in order to allow for independent resource scaling and thus, better hardware resource utilization. While disaggregation of hard-drives storage is a common practice, NVMe-SSD (i.e., PCIe-based SSD) disaggregation is considered more challenging. This is because SSDs are significantly faster than hard drives, so the latency overheads (due to both network and CPU processing) as well as the extra compute cycles needed for the offloading stack become much more pronounced.

In this work we characterize the overheads of NVMe-SSD disaggregation. We show that NVMe-over-Fabrics (NVMf) - a recently-released remote storage protocol specification - reduces the overheads of remote access to a bare minimum, thus greatly increasing the cost-efficiency of Flash disaggregation. Specifically, while recent work showed that SSD storage disaggregation via iSCSI degrades application-level throughput by 20%, we report on negligible performance degradation with NVMf - both when using stress-tests as well as with a more-realistic KV-store workload.

DOWNLOAD

2016

Understanding performance of I/O intensive containerized applications for NVMe SSDs

Abstract:
Our cloud-based IT world is founded on hypervisors and containers. Containers are becoming an important cornerstone, which is increasingly used day-by-day. Among different available frameworks, docker has become one of the major adoptees to use containerized platform in data centers and enterprise servers, due to its ease of deploying and scaling. Further more, the performance benefits of a lightweight container platform can be leveraged even more with a fast back-end storage like high performance SSDs. However, increase in number of simultaneously operating docker containers may not guarantee an aggregated performance improvement due to saturation. Thus, understanding performance bottleneck in a multi-tenancy docker environment is critically important to maintain application level fairness and perform better resource management.

In this paper, we characterize the performance of persistent storage option (through data volume) for I/O intensive, dockerized applications. Our work investigates the impact on performance with increasing number of simultaneous docker containers in different workload environments. We provide, first of its kind study of I/O intensive containerized applications operating with NVMe SSDs. We show that 1) a six times better application throughput can be obtained, just by wise selection of number of containerized instances compared to single instance; and 2) for multiple application containers running simultaneously, an application throughput may degrade upto 50% compared to a stand-alone applications throughput, if good choice of application and workload is not made. We then propose novel design guidelines for an optimal and fair operation of both homogeneous and heterogeneous environments mixed with different applications and workloads.

Full Citation: http://ieeexplore.ieee.org/document/7820650/

DOWNLOAD

2015

System-Level Characterization of Datacenter Applications

In recent years, a number of benchmark suites have been created for the "Big Data" domain, and a number of such applications fit the client-server paradigm. A large volume of recent literature in characterizing "Big Data" applications have largely focused on two extremes of the characterization spectrum. On one hand, multiple studies have focused on client-side performance. These involve fine-tuning serverside parameters for an application to get the best client-side performance. On the other extreme, characterization focuses on picking one set of client-side parameters and then reporting the server microarchitectural statistics under those assumptions. While the two ends of the spectrum present interesting results, this paper argues that they are not enough, and in some cases, undesirable, to drive system-wide architectural decisions in datacenter design. Abstract:

This paper shows that for the purposes of designing an efficient datacenter, detailed microarchitectural characterization of "Big Data" applications is an overkill. It identifies four main system-level macro-architectural features and shows that these features are more representative of an application's system level behavior. To this end, a number of datacenter applications from a variety of benchmark suites are evaluated and classified into these previously identified macro-architectural features. Based on this analysis, the paper further shows that each application class will benefit from a very different server configuration leading to a highly efficient, cost-effective datacenter.

DOWNLOAD

2015

Performance Characterization of Hyperscale Applications on on NVMe SSDs

Abstract:
The storage subsystem has undergone tremendous innovation in order to keep up with the ever-increasing demand for throughput. NVMe based SSDs are the latest development in this domain, delivering unprecedented performance in terms of both latency and peak bandwidth. Given their superior performance, NVMe drives are expected to be particularly beneficial for I/O intensive applications in datacenter installations. In this paper we identify and analyze the different factors leading to the better performance of NVMe SSDs. Then, using databases as the prominent use-case, we show how these would translate into real-world benefits. We evaluate both a relational database (MySQL) and a NoSQL database (Cassandra) and demonstrate significant performance gains over best-in-class enterprise SATA SSDs: from 3.5 × for TPC-C and up to 8.5 × for Cassandra.

DOWNLOAD

2015

Performance analysis of NVMe SSDs and their implication on real world databases

Abstract:
The storage subsystem has undergone tremendous innovation in order to keep up with the ever-increasing demand for throughput. Non Volatile Memory Express (NVMe) based solid state devices are the latest development in this domain, delivering unprecedented performance in terms of latency and peak bandwidth. NVMe drives are expected to be particularly beneficial for I/O intensive applications, with databases being one of the prominent use-cases.

This paper provides the first, in-depth performance analysis of NVMe drives. Combining driver instrumentation with system monitoring tools, we present a breakdown of access times for I/O requests throughout the entire system. Furthermore, we present a detailed, quantitative analysis of all the factors contributing to the low-latency, high-throughput characteristics of NVMe drives, including the system software stack. Lastly, we characterize the performance of multiple cloud databases (both relational and NoSQL) on state-of-the-art NVMe drives, and compare that to their performance on enterprise-class SATA-based SSDs. We show that NVMe-backed database applications deliver up to 8x superior client-side performance over enterprise-class, SATAbased SSDs.

DOWNLOAD

Product Overviews