As of September 7, 2018, Huawei V5 series servers have topped a total of 121 SPEC benchmark tests. Powered by Intel^® Xeon^® Scalable processors, Huawei servers are recognized in the industry for their performance, quality, and power efficiency. As published on the SPEC official website, the winning results of these 121 SPEC benchmark tests endorse the competitiveness of Huawei servers.

The test results of SPEC, an authoritative third-party comprehensive evaluation institute in the industry, have been an important reference for enterprise customers to select servers. Huawei servers top 121 SPEC tests that cover a wide spectrum of SPEC tests, including SPECint^®_base2006, SPECfp^®_base2006, SPECfp^®_rate_base2006, SPECint^®_base2017, SPECfp^®_base2017, SPECint^®_rate_base2017, SPECfp^®_rate_base2017, and SPECpower-ssj2008. These tests focus on some of the most important indicators for measuring server performance and energy consumption, including integer computing performance, floating-point computing performance, and performance per watt.

• Integer Computing Performance

Nine Huawei server models, including FusionServer 2288H V5/1288H V5/CH121 V5/5288 V5/CH242 V5/XH321 V5/2488 V5/2488H V5/5885H V5, ranked No.1 in 58 performance tests of SPECint^®_base2006, SPECint^®_base2017, and SPECint^®_rate_base2017.

• Floating-Point Computing Performance

Eight Huawei server models, including FusionServer 2288H V5/1288H V5/CH121 V5/5288 V5/CH242 V5/2488H V5/5885H V5/KunLun 9008 V5, ranked No.1 in 62 performance tests of SPECfp^®_rate_base2006, SPECfp^®_base2017, and SPECfp^®_rate_base2017.

• Performance per Watt

Huawei FusionServer 2288H V5 ranked No.1 in the energy efficiency test of SPECpower-ssj2008.

Server performance and energy consumption depend on complex factors. The following uncovers Huawei servers’ secrets of winning 121 SPEC benchmark tests.

1. Temperature Largely Affects Processor Performance, and Heat Dissipation Holds the Key

Huawei servers adopt an efficient end-to-end (E2E) heat dissipation design that extends from air intake vents to air exhaust vents. Aided by multiple temperature sensors, the servers can adjust fan module speeds in real time according to ambient temperatures. Huawei servers boast a string of unique designs, such as cellular air intake vents with a porosity rate of 75%, counter-rotating fan modules that improve the wind speed by 30%, 3D VC heat sinks that feature integrated pouring technique and improve heat dissipation efficiency by 3%, flow-dividing air ducts that improve heat dissipation efficiency by 9%, and independent air channels designed for I/O modules.

2. High-Quality Signal Links Ensure Smooth Signal Transmission

Huawei servers are designed by three dimensions. Components are deployed in the X, Y, and Z spaces, utilizing physical space to the full and reducing interference between signal links. In terms of simulation, optimization, and processing techniques, Huawei performs three-dimensional full-sized modeling and simulation tests and optimizations on BGA ball vias, capacitors, and connectors to control signal fluctuation in a proper range and improve the impedance consistency of links. Huawei has also introduced the signal digital eye pattern detection technology into the server production. This technology can quickly identify boards with poor link signal quality, which is difficult to detect by common methods.

3. In-Depth Hardware/Software Optimization from Chips to BIOS

Huawei-developed BMC chips and BIOS source code can be integrated with hardware for in-depth optimization. For example, to maximize memory performance, parameters like SNC (Sub NUMA Clustering), IMC (Integrated Memory Controller) Interleaving, and XPT (Xtended Prediction Table) Prefetch can be adjusted to improve memory bandwidth. Similar in-depth optimization can also be performed on processors and hard drives to achieve optimal performance. 

4. Dynamic Energy Management Technology for Energy Saving

Huawei’s patented Dynamic Energy Management Technology (DEMT) is the key to intelligent power consumption management. According to processor load and ambient temperatures, the technology intelligently and dynamically adjusts processor, memory, hard drive power consumption and fan module speeds using functions such as frequency modulation and hibernation, multiple-input multiple-output (MIMO) speed adjustment, proportional-integral-derivative (PID), and Deep Sleeping Technology (DTS 2.0). This cutting-edge technology reduces server power consumption without compromising performance.

Huawei V5 series servers include various product models, including FusionServer 2- to 4-socket rack servers, high-density servers, blade servers, and KunLun mission critical servers. In pursuit of efficiency and agility, these servers are positioned for future-oriented complex service workloads, such as cloud computing, big data, and artificial intelligence. In response to customers’ core requirements amid digital and intelligent transformation, in-depth innovations have been made on servers from structural design to production and manufacturing, and from underlying hardware to upper-layer management software. All these help Huawei servers win the first place of 121 SPEC benchmark tests. In addition to strong performance, Huawei V5 servers boast five intelligent management features: intelligent software management, intelligent automatic deployment, intelligent fault prediction and diagnosis, intelligent power consumption management, and unified asset management. These features help reduce the OPEX for large-scale deployment by 5–15%, making Huawei a preferred IT infrastructure partner for enterprise customers to implement digital and intelligent transformation.

▲ Huawei V5 server family

 

The post The Secret Behind Huawei’s Top Rank in 121 SPEC Benchmarks appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

By Wang Jiaxin from Huawei

On October 11 2017, the Low Carbon Vehicle Partnership (LowCVP) announced BYD Auto as the 2017 Low Carbon Champion. BYD Auto won the “Low Carbon Heavy Duty Vehicle of the Year Award” and, in partnership with Transport for London (TfL) and Go-Ahead London, won the “Grand Prix” – or winner-of-winners-award – for launching London’s first all-electric bus routes (the 507 and 521). These awards specifically noted BYDs incredible efficiency in terms of delivering the vehicles, operational capacity, and building infrastructure for the London project.

Founded in 1995, BYD engages in IT-related and automobile business for traditional fuel-powered vehicles and vehicles powered by new types of energy sources. With 30 production centers established around the globe, BYD is the only vendor worldwide that owns power battery and vehicle technologies in the new energy field.

Stretched ERP System of Unclear Causes

The ERP system is a core system that supports BYD’s three product lines (IT, automobile, and new energy). The original legacy ERP system could not deliver the necessary stability and performance to meet growing business expansion requirements, for example, the number of employees more than doubling, factory orders growing by 10% every quarter (reaching 185,000), and concurrent connections to the ERP system exceeding 60,000. To make matters worse, the ERP system took more than 1 hour to query financial cost control-related reports and the system would occasionally breakdown during monthly settlements, severely affecting the operating of the production line.

Troubleshooting with Huawei

As a leading global ICT solutions provider, Huawei offers end-to-end solutions that cover networks, cloud computing, servers, databases, and storage. BYD decided to team up with Huawei, a company with extensive experience in IT solutions, to analyze server memory, CPU utilization, networks, and disk response times. The team discovered that over 200 GB of storage data that was bottlenecking service access performance was generated in the ERP’s production system every month. In addition, the team found that the I/O latency was reaching 50 ms on the live network.

The original and outdated HDD-based storage devices in the ERP system delivered poor access performance. BYD wanted to improve its business performance by 100% without changing the existing ERP system architecture. Smoothly migrating applications without disrupting application systems was a challenge for Huawei.

Huawei OceanStor Dorado leads to a three-fold increase in BYD ERP performance

Based on the requirements and analysis results of BYD’s live network, Huawei provided the OceanStor Dorado all-flash acceleration solution which leverages all-flash storage. This allows the seamless replacement of existing HDDs, accelerating read and write I/Os of the ERP and BW systems and improving system efficiency. This solution does not change the architecture, database, host, ERP software, or O&M mode of the live network. The minimal change translates to minimal risks.

To prevent system upgrades from affecting services, the Huawei data migration team provided a set of data mirroring migration solutions by using host volume management software. These solutions migrated data from legacy storage to OceanStor Dorado all-flash storage without affecting ERP services. In addition, Huawei recommended that BYD run development and test systems, such as sandbox, on the legacy storage system to ensure return on investment (ROI).

After the ERP system switches to Huawei OceanStor Dorado, the latency is reduced to less than 1 ms. With the query condition design and upper-layer platform of ERP, the data extraction time for BW data warehouses is reduced by 66%. Additionally, the report output speed of the MM, PP, SD, and FICO modules sees an increase of more than 200%. The time required to query material documents is reduced from 14 minutes to 4 minutes.

Legacy HDDs are configured with RAID 10 groups and both SAS disks and SSDs are deployed. To meet service capacity and performance requirements under these conditions, 40 U cabinets are required. Huawei OceanStor Dorado all-flash storage uses RAID 6 groups and all-SSD configurations, requiring only 20 U cabinets. This saves 50% on cabinet space and O&M manpower costs, while significantly reducing electricity and cooling costs. In addition, the estimated three-year OPEX can be slashed by more than 70%, ensuring competitive advantages for all-flash storage compared to traditional HDD-based storage.

Huawei OceanStor Dorado has now been running stably for more than half a year in the ERP system on BYD’s live network. BYD has seen noticeable improvements to service performance and reductions in maintenance costs. The IT system reconstruction has surpassed all of BYD’s expectations.

Inspired by the successful reconstruction of the ERP system, BYD once again chose Huawei’s Dorado all-flash storage to upgrade and revamp its DMS system. Dorado’s superb performance and robust stability have once again impressed BYD. It is expected that BYD and Huawei will further their cooperation through the smart manufacturing of “Made in China 2025”.

The post BYD Auto Becomes a Leader in New Energy with All-Flash Storage appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

At the recently concluded Jakarta Asian Games, Chinese athletes struggled and maintained their medal rankings with 132 gold, 92 silver and 65 copper. This national strength is also reflected in the enterprise IT field, where professional testing organizations set the scene for IT vendors to demonstrate their pursuit of the ultimate user experience.

SAP is an industry leader in software. In addition to commercial software for enterprises, SAP provides benchmark tests built on online analytical processing (OLAP) and online transaction processing (OLTP) applications for SAP-certified hardware.

The KunLun 9008 V5 is Huawei’s next-generation mission critical server. It recently topped the SAP Sales and Distribution (SD) benchmark for OLTP applications with an SAP Application Performance Standard (SAPS) score of 587,180. This is not KunLun’s first time as an award winner – in November 2017, its predecessor KunLun 9008 scored a leading 425,550 on the same benchmark.

The SAP SD benchmark measures the performance of OLTP applications on servers. Testing involves the installation of service applications and databases on the same physical server, and the benchmark test tool on the client. Results are standardized as SAPS scores for the SAP SD application module, where 100 SAPS is defined as 2,000 fully business processed order line items per hour. Fully business processed means the full business process of an order line item: creating the order, creating a delivery note for the order, displaying the order, changing the delivery, posting a goods issue, listing orders, and creating an invoice with a blank invoice template. In technical terms, 100 SAPS is achieved by processing 6,000 dialog steps (screen changes) or 2,000 postings per hour in the SD benchmark.

Powered by the latest Intel^® Xeon^® Scalable processors, Huawei KunLun 9008 V5 provides up to 224 processor cores and 12 TB of memory. The KunLun 9008 V5 server tested in the latest SAP SD benchmark was configured with eight Intel^® Xeon^® Platinum 8180M Scalable processors and 3 TB of memory. It performed as follows:

Source: SAP official website

KunLun is not just about excellent performance. The x86 server with the highest reliability, KunLun comes with cutting-edge RAS 2.0 technology and BMC management chips to support the open x86 ecosystem and the reliability of UNIX servers. Services remain uninterrupted in mission-critical scenarios such as core databases, key applications, and in-memory computing databases.

Solid reliability: KunLun’s early fault warning, automatic fault isolation, and online replacement of faulty components without interrupting services make it the industry’s only server with CPU and memory module hot swap.

Flexible integration: KunLun pioneers physical and logical partitioning and hybrid deployment of both physical and logical partitions. With a processor core as its basic unit, a logical partition is deployed on multiple physical partitions for on-demand resource allocation and maximum utilization.

Open ecosystem: KunLun supports the operating systems, virtualization software, database systems, and applications that are mainstream in the industry. With partners from the finance, government, telecom, retail, and manufacturing industries, Huawei has built KunLun on x86 architecture with standardized software and hardware deployment and O&M, so customers reduce costs.

Huawei partners SAP to provide single-node (192 GB to 16 TB) or cluster (up to 96 TB) solutions for SAP HANA. SAP-certified and KunLun-based, these solutions have been key to the digital transformation of more than a thousand enterprises from China, Russia, Spain, Germany, Italy, Algeria, Saudi Arabia, India, and Singapore.

When it comes to mission-critical applications, enterprises want reliable and powerful IT infrastructure. This field is thus of strategic significance for IT vendors demonstrating their technical strength. KunLun’s repeat victory with the SAP SD benchmark bolsters Huawei’s confidence in providing competitive IT products and solutions for global enterprise customers.

The post KunLun Mission Critical Server Defends SAP Benchmark Championship appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

By Li Wuxian from Huawei

Enterprise data is stored in professional storage. For large enterprises comprising many data centers and massive volumes of data, storage is often a pain point. Storage devices need to both provide instant access and take up less floor space, all at a lower price than other alternatives. Storage device replacement is required when devices are out of warranty, end of life, or encounter insufficient performance and capacity.

Replacing storage devices involves business integration and data migration. Data migration does not mean simply moving data from one place to another. Instead it involves data security, compatibility, migration efficiency, outage time, as well as considerations of its impact on business. Currently, there are three main types of data migration: at the application layer, at the host layer, and at the storage layer. Because of the many different aspects surrounding storage and data migration, it is no wonder that it is the focus of many industry giants. From a customer standpoint, enterprises need to effectively judge how data migration will impact their business and services. Further, they need to find ways to both enable online data takeover and migration, and ensure a quick migration speed, as well as choose a solution that delivers an easy-to-use interface for future business expansion and integration. During this process, if any financial challenges arise, consulting with a firm rescue skilled or insolvency practitioner can offer the guidance needed to navigate complexities and maintain financial stability while focusing on growth and development.

Analysis of Mainstream Data Migration Solutions

Data migration at the application layer: A major migration method based on database and virtual machines (VMs) functions at the application layer, this solution is used in cross-platform and cross-media migration scenarios and scenarios with short downtime. It features simple access to the target storage and little impact on the source network topology. In some scenarios, migration and replacement tolerate short service interruption.

Because this method of storage requires host resources during data migration, it can create huge pressure on service systems. It is often associated with low migration speeds, sometimes even less than 200 MB/s. In addition, has application-layer migration solutions have restrictions on application configuration. For example, Oracle ASM Rebalance supports only Oracle ASM to manage disks, and VMware Storage vMotion does not support online migration in some scenarios (VM shared disks and RDM disks, etc.).

Data migration at the host layer: implemented by the volume management software. Common Veritas Volume Manager (VxVM) of a third party and Logical Volume Manager (LVM) of AIX, HP-UX, and Linux are used to migrate data online. Host-layer migration solutions ensure that services are not interrupted during migration, satisfying storage space consolidation and cross-media migration demands. However, the solutions may provide slow migration speeds of less than 100 MB/s, as well as large consumption of host CPU and memory resources. In addition, manual inputting of command lines is required during the migration, which is prone to errors.

Data migration at the storage layer: Utilizes the built-in functions of storage devices to migrate data at the storage layer. These functions are as follows:

LUN copy and remote replication: used for migration between homogeneous storage, where the migration speed of LUN copy can reach up to 500 MB/s.

Remote replication: used for long-distance migration, where its online migration speed is about 100 MB/s on average.

Heterogeneous virtualization: uses storage virtualization functions to take over data from heterogeneous devices and then migrate the data to the source end. Its migration speed is up to 500 MB/s and third-party storage devices are supported.

Data migration at the storage layer features fast migration speeds and does not impact host system performance. Likewise, there is no vendor lock-in, allowing users to migrate to and from third-party storage devices. However, LUN copy is offline, and the switchover of remote replication service takes a long time. Heterogeneous virtualization is insufficient, and the takeover requires the device to suspend.

Storage-layer migration solutions need to provide non-disruptive data migration processes, and support online takeover migration.

Based on these demands, Huawei launched OceanStor V5 Hybrid Flash Storage, OceanStor F V5, and OceanStor Dorado All Flash Storage series. These are business-unaware migration solutions with zero downtime perfectly designed to optimize heterogeneous virtualization, implementing online takeover and migration of data at the storage layer.

Huawei’s Heterogeneous Virtualization Data Migration Solution

OceanStor V5 Hybrid Flash Storage, OceanStor F V5, and OceanStor Dorado All Flash Storage series adopt SmartVirtualization and SmartMigration in heterogeneous virtualization migration scenarios. These functions migrate data in the source array to the target array at the storage layer. Users can use SmartVirtualization to take over the source array (third-party storage) online, and then migrate the data online by using SmartMigration. After the migration is complete, users can remove the source storage from the live network.

Implementation Principles of Huawei’s Heterogeneous Virtualization Migration Solutions

Customers can move data from third-party storage to Huawei’s storage External LUNs on third-party storage devices to store actual data. These Huawei devices are commonly known as eDevLUNs. An eDevLUN consists of a Meta Volume and a Data Volume.

The Meta Volume is used to manage the data organization of eDevLUNs. The storage space of Meta Volumes is provided by Huawei’s storage devices, and occupies little space.

A Data Volume is a logical abstract object of external LUN data. An external LUN provides all data storage space. eDevLUNs created on Huawei’s storage systems have one-to-one mapping relationships with external LUNs. Application servers manage data on external LUNs through data reads and writes on eDevLUNs.

In a heterogeneous virtualization solution, LUNs mapped to source arrays are taken over during the initial migration phase, and LUNs of third-party storage are initialized into eDevLUNs that can be managed by the target arrays under the condition that no data is lost. LUNs are migrated using value-added features of heterogeneous virtualization, which enables data to be rapidly migrated between the source and target arrays. Heterogeneous takeover can be performed online, and the services are non-disruptive during data replication.

Huawei’s Online Migration Technologies in Heterogeneous Virtualization Scenarios

Compared with takeovers in traditional heterogeneous virtualization scenarios, Huawei’s takeover can encapsulate the source LUNs into eDevLUNs without interrupting host services. Huawei’s storage maps the eDevLUNs to the host so that the hosts’ multi-pathing software can automatically aggregate the physical paths of the source LUNs and eDevLUNs. The software runs host multi-pathing commands to delete the corresponding physical links of the source LUNs and switch all I/Os to the link to implement online takeover with zero downtime.

Process of Online Takeover and Migration

1. Connect the source arrays to the target arrays, and map the LUNs to be migrated to Huawei’s disk arrays.
2. The Huawei’s disk arrays encapsulate the LUNs mapped to the source arrays into eDevLUNs that are then mapped to the source host.
3. Use multi-pathing software on the host system to aggregate physical links between the source LUNs and eDevLUNs and delete the physical links of the source LUNs to complete online takeover.
4. After the online takeover, use SmartMigration to migrate data from the source LUNs to Huawei’s disk arrays to implement online data migration.

Advantages of Huawei’s Online Migration in Heterogeneous Virtualization Scenarios

• Mainstream operating systems can be used to migrate data between SAN storage arrays from different vendors.
• During migration, data is replicated on storage arrays, which does not occupy CPU or memory resources of the host, imposing zero impact on business.
• Services are not interrupted during array takeover and data migration
• Data migration speed can reach up to 500 MB/s.
• Data can be migrated conveniently and quickly without installing any extra software or agent in the existing environment.
• Migration Director is a professional migration tool that supports automatic operation on an efficient graphical user interface.

The post Huawei’s Business Integration Solution: A Revolutionary Optimal Data Migration Method appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

By Liu Lin from Huawei

Public cloud services are developing rapidly all around the world. Enterprise data centers are quickly evolving to the cloud, leaving traditional storage methods obsolete in the 5G era. Reports have shown that 68% of enterprises in North America have started to use cloud services to various degrees. Companies in other parts of the world are reported to be trying out the new cloud-based service models.

What has enterprises around the world so excited about the new service models?

In terms of improving overall business production for enterprises in many different industries, public cloud business models are a revolutionary evolution. The likes of Huawei’s cloud family have never been seen before. They are easy to use and reduce TCO, and the ease of use is provided with on-demand and quick resource acquisition capabilities. Enterprises are always looking for ways to save on TCO through adoption of new applications and techs. According to statistics, the costs for storing data in self-purchased storage are 19 to 30 times higher than those for leasing public cloud storage. Considering the costs, it’s no wonder more and more enterprises are using cloud services.

Large cloud service providers can handle just about any workload for organizations of all sizes. However, enterprises are still somewhat reserved about placing their mission-critical services and data on the cloud due to the security requirements levied and high performance requirements. To address any reservations that enterprises may have, many cloud service providers deliver high security guarantees to prove to potential customers that migrating mission-critical services to the public cloud is safe, quick, and easy.

Dedicated Enterprise Storage Service (DESS) migrates mission-critical services to the cloud without making any forklift upgrades

HUAWEI CLOUD has a proven track record of helping many move their services to the cloud, even in the most demanding of scenarios. With HUAWEI CLOUD, security and extreme performance are no longer blocking mission-critical services from moving to the cloud.

DESS uses Huawei all-flash and hybrid-flash storage to build dedicated storage services dedicated to enterprises. Oriented to enterprises’ key applications, such as Oracle RAC and SAP HANA TDI, it provides the same superb performance and reliability as enterprise private cloud, simplifying service deployment on public cloud for enterprise users.

DESS users can exclusively use physical storage resources in HUAWEI CLOUD to ensure that services are stably running on isolated storage devices. In a typical database performance model, databases can obtain a stable latency of 0.5 ms and a read/write capability of 3000 IOPS/TB. Thanks to the robust reliability of enterprise-class storage, users can use the same high-reliability solution as that used in self-owned data centers, with 99.99999% of data durability.

Since its release in September 2017, DESS has been used by many large enterprises and government agencies to successfully deploy their key applications on the cloud. For example, after migrating the ERP system to DESS, a large-sized private healthcare company has reduced the time needed to generate financial statements from several days to just several hours, helping to improve enterprise business analysis and decision-making efficiency by dozens of times.

STaaS solution centrally manages resources on- and off-premises

For customers who have some of their services running on the cloud, Huawei also provides the Storage as a Service (STaaS) solution to facilitate storage resource management and data flow on- and off-premises. The STaaS solution can manage various storage devices in the on-site data center and access to HUAWEI CLOUD. Enterprises can manage on- and off-premises resources over a single pane of glass, eliminating the troubles in switching back and forth between interfaces.

The rich selection in the service catalog enables high degree of automation in provisioning of storage resources at both on-site and cloud-based data centers. Even such operations as volume creation, file sharing, creation of active-active relationships, volume backups, volume replications are all easily handled with time-saving automations.

The results speak for themselves. KIO, the largest ISP in Mexico, uses the STaaS solution to centrally manage and schedule storage resources in multiple data centers, reducing manual operations and improving work efficiency by over 50%.

The impressive on- and off-premises resource management and automated service provisioning capabilities have won over customer after customer in finance, telecom, public services, retail, and other sectors. What’s more, after intense vetting and proven track record with zero incidence after deployment, Huawei has since become a brand that you can trust. Therefore it’s no surprise the Huawei STaaS solution has become the compelling choice in hybrid cloud environments able to carry the mission-critical services for organizations of all sizes.

 

 

The post Moving Your Mission-Critical Services to Cloud Just Got Easier! appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

By Zhao Yanling from Huawei

The ESG lab is a well-known IT consulting organization based in the United States. It performs IT analyses, research, and verification, and is renowned for its independent, rigorous, and practical laboratory tests.

In September 2017, the ESG lab tested the performance, reliability, cost-effectiveness, and total cost of ownership (TCO) of Huawei OceanStor Dorado V3. The ESG set up a real service environment and tested the performance and reliability of Huawei all-flash storage for applications such as VMs, databases, and mailboxes.

The following figure shows the test environment used in the ESG lab. The VMware vSphere 6.0 virtualization infrastructure is deployed on 10 Huawei FusionServer RH series rack servers, with two dual-controller Huawei OceanStor Dorado5000 V3 storage systems connected to the back-end.

According to the ESG lab test data, Huawei Dorado V3 all-flash storage delivers a consistent latency of 0.3 ms under the pressure of 5000 Exchange users, 1000 VMware Horizon virtual desktops, and hundreds of Oracle users. Even if storage devices are faulty, upper-layer applications can automatically recover within 3 seconds by using the active-active solution, ensuring service continuity. In addition, the data reduction technology reduces the five-year total cost of ownership (TCO) by 75%, making Huawei Dorado V3 an ideal choice for mission-critical workloads.

Stable High Performance

The ESG lab has verified that the OceanStor Dorado V3 can integrate the most challenging workloads and mission-critical tasks into a high-performance and high-availability platform. When 5000 Exchange users, 1000 VMware Horizon virtual desktops, and hundreds of Oracle users are integrated into one OceanStor Dorado5000 V3 system, the integration environment has an average response time of 0.3 ms for these services.

The following figure shows the test results of five thousand 1 GB mailboxes and 9.6 TB Oracle OLTP databases. In this case, the system supports about 105,000 continuous IOPS, while maintaining an average response time of just 0.3 milliseconds.

In the next test, personnel simulated a boot storm of 1000 virtual desktops. The I/O soared to about 160,000 IOPS in approximately one minute. At this time, the peak response time increased but remained lower than 0.5 ms. Shortly after this, the two indicators returned to the same level as before the boot storm.

In addition, the ESG lab created multiple snapshots and then rolled these back, which did not affect system performance. Deduplication and compression are always in the enabled state, and the deduplication and compression ratio is 7:1 for the three types of hybrid workloads. Together with Thin Provisioning, this reduces the total space required by 14:1.

High Availability

Huawei OceanStor Dorado V3 all-flash platform ensures high availability and stable performance during planned maintenance and unexpected interruptions. The HyperMetro feature provides a 100-kilometer, gateway-free, active-active solution to ensure service continuity. RAID-TP technology ensures services continue to run stably even if three disks in a single disk array fail.

In the ESG lab, HyperMetro is first tested in a campus environment. Two OceanStor Dorado V3 storage systems are configured with HyperMetro and run the Oracle RAC database to simulate OLTP transaction database traffic. The total traffic generated by the two servers is 50000 IOPS, the processing volume of a single disk array is 25000 IOPS, and the average I/O response time is 0.25 ms.

The ESG simulated the failure of array 1 by removing its power supply, reducing the total IOPS of array 2 to 0. After about 3 seconds, the total IOPS of array 2 was restored to provide services for all requests in the cluster. A total of 50,000 IOPS are supported, and during the switchover the maximum latency on the controller of array 2 was 4 ms and 8 ms. The latency was reduced to 0.25 ms after 3 seconds.

The ESG also tested the performance of the Dorado RAID TP technology when three disks are faulty. Three SSDs were removed from the disk array at intervals of one to two minutes to simulate the simultaneous failure of three SSDs. The disk array was still available after three SSDs were shown to be faulty without significantly reducing performance.

The RAID-TP reconstruction time was also tested. The ESG lab removed an SSD from a disk array that was configured with nine SSDs and 7.4 TB volumes. It took just nine minutes for the RAID-TP to reconstruct the 7.4 TB volume. There is a 7.4 TB capacity and there are nine SSDs, while the reconstruction data volume of each SSD is about 822 GB. Therefore, the reconstruction time of 1 TB of data was only 11 minutes.

TCO

The ESG lab simulated the expected storage TCO of a company that required a high availability hybrid service environment. The OceanStor Dorado5000 V3 has an obvious advantage in this area when compared with Huawei’s hybrid storage arrays and the first-generation all-flash storage devices of other mainstream vendors. The five-year TCO was shown to be reduced by more than 75%.

The ESG lab acknowledges that Huawei OceanStor Dorado V3 provides stable high performance and high reliability for its users. Huawei OceanStor V3 perfectly adapts to high-performance service applications and can run in a virtualization environment with high performance requirements.

The post ESG Test Report: Huawei OceanStor Dorado V3 Delivers Excellent Performance for Mainstream All-Flash Applications appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—

By Wang Taiping from Huawei

Purpose-built for mission-critical businesses, OceanStor Dorado V3 all flash storage system delivers high performance, reliability, and efficiency augmented by FlashLink technology.

• Innovative FlashLink technology: enables 0.5 ms stable latency and 3-fold better storage performance than that of traditional storage.
• HyperMetro gateway-free active-active design: provides smooth business migration to DCs in geo-redundant mode and solution-level 99.9999% availability, with RPO = 0 and RTO ≈ 0.
• Inline deduplication and compression function: helps Huawei storage claim 3:1 data reduction guarantee (depending on application and configuration) and reduce TCO by 75%.

Huawei OceanStor Dorado V3 fully satisfies the high performance, reliability, and efficiency requirements of mission-critical database services and virtualization efforts, smoothing the way for customers in finance, manufacturing, telecom, and other sectors in the all-flash era.

Database acceleration scenario (Oracle OLTP)

Oracle database hardware applications are the most widely used series by large enterprises. However, today’s information explosion poses great challenges to the performance and costs of Oracle systems.

1. Stacking of devices

Large e-commerce and banking systems process more than 100 million transactions per day. Core business systems perform in excess of 200,000 IOPS while keeping the latency within only 1 to 3 ms. However, traditional storage systems simply stack more devices to achieve increases in performance.

2. Business continuity

Oracle OLTP carries mission-critical businesses for enterprises, remaining unchanged pursuit of business continuity.

3. Increasing costs for Oracle licensing

Oracle licensing costs account for a significant portion of the O&M costs in enterprise IT systems. For example, a database with 10 TB of capacity could cost an enterprise as much as 50% of its yearly maintenance budget. As data volumes increase, the annual licensing costs would also continue to rise. Reducing the licensing costs for Oracle applications is a top concern for enterprises.

Not having enough storage performance could stunt the growth of the enterprise, while not having disaster recovery capabilities in place could jeopardize the entire operation.

Huawei Dorado provides high performance, robust continuity assurances, and reduced TCO, ensuring the mission-critical OLTP services of the enterprise remain online while adding new degrees of efficiency in operations.

Huawei all-flash Oracle database (OLTP) solution

The intense performance of the Huawei all flash storage solution powered on OceanStor Dorado V3 doubles online transaction efficiency while keeping latency down to a stable 0.5 ms, minimizing wait times, and improving resource utilization of storage and server assets. Many enterprises have seen a 50% reduction in Oracle licensing costs. The scale-out layout allows for easy, linear expansion. Capacity increases are predictable and scale-outs can be implemented in nearly no time, allaying enterprise concerns over the unpredictability factors in service growth. The Huawei offerings going into the solution provide a rich catalog of data protection services, ensuring that mission-critical applications remain online 24/7.

Highlights

1. More than 1.5 times faster performance than that of other all-flash storage

After tested by using the simulated system for Oracle database online transaction, a dual-controller OceanStor Dorado V3 system can provide stable performance of 210, 000 IOPS at 0.5 ms latency, 10 times higher than that of traditional storage and 1.5 times than that of XtremIO X2.

2. Complete DR solution

OceanStor Dorado V3 with the active-active HyperMetro gateway-free feature enabled achieves 99.9999% high availability for mission-critical businesses. The gateway-free attribute of the solution reduces the possible points of node failure, simplifies the networking, minimizes frequency of error from manual operations, and improves overall system reliability. The active-active architecture is able to balance loads and execute seamless service switching.

3. Lower Oracle licensing costs

The ultra-low latency on Dorado V3 reduces the wait time for CPU I/Os and effectively lowers CPU usage. Under the same workload, using less CPU resources helps increase processing speeds and lower Oracle licensing fees. As the figure shows, compared to traditional storage, OceanStor Dorado V3 storage reduces the CPU usage by about 50%.

VDI scenario (VMware)

Terminal management and delivery services are rather time-intensive and costly in traditional office layouts dominated by PCs. Virtual desktop infrastructures (VDIs) have become widely adopted in enterprise IT systems thanks to the ability to deliver, protect, and manage Windows-based or Linux-based desktops and applications on premise or in the cloud all while controlling costs. Adding to the luster, users are able to work from any location and at any time they chose. However, the availability, performance, and maintenance requirements on centralized deployment of VDI systems are much higher than that of the traditional PC environment.

1. High costs

The VMware desktop solution claims the annual costs in typical use cases (30 GB system disks, 10 GB data disks) is $699 USD for skilled users, which is higher than the traditional office solution. The investment into storage in the VDI solution accounts for 25 to 40% of the total, a big cost consideration to adopting such a layout.

2. Time-consuming O&M

Deploying a VDI environment for 2000 users takes over 20 hours with traditional storage and it takes considerable time to perform expansions. When emergency situations require patching, it can take over 10 hours to install the patches and perform the restarts for 2000 desktops, causing serious impact to daily operations.

3. User experience gap between virtual desktops and PCs

The traditional VDI solution delivers response time of about 5 to 10 seconds to open a slide, Word, Excel, and other files.

Compared to traditional storage devices, the Huawei OceanStor Dorado VDI solution provides a new balance among high performance, service continuity, and efficiency.

Huawei VDI solution

Huawei OceanStor Dorado V3 all-flash storage provides industry-leading performance and low latency, ensuring quick maintenance of VDI systems and good customer experience. The rich store of DR solutions available with the Dorado fully satisfies different reliability requirements. With the inline data reduction function enabled, TCO is significantly reduced in VDI environments.

Highlights

The following testing indicates the experience differences between Huawei OceanStor Dorado V3 and traditional storage arrays in common VDI scenarios.

Comparison of the best practice data of mainstream products and test results of Huawei solution is shown as follows:

1. More than 30% lower cost per desktop

Comparing the single desktop cost among OceanStor Dorado V3, XtremIO, and traditional storage, it can be seen that OceanStor Dorado V3 outperforms XtremIO with the same 5:1 deduplication and compression ratio, while traditional storage did not enable this function. More than 30% of costs can be reduced when over 1000 desktops are operating.

2. High O&M efficiency

OceanStor Dorado V3 storage reduces the time to deploy 2000 desktops from 20 hours to 8 hours compared to traditional storage and lowers the startup storm time by 75% (60 minutes to 15 minutes).

3. Better user experience

Compared to traditional storage, the response time to open office applications is shortened by 50%.

Scalable VDI storage solution

Server virtualization scenario (VMware)

Storage is the biggest bottleneck for a system to achieve centralized business deployment and maximum resource utilization. However, the server virtualization technology can dramatically improve the IT efficiency and elasticity for enterprises. Huawei all-flash storage can cope high performance challenges in server virtualization scenarios, enabling enterprises to avail themselves of the advantages.

1. Business integration

In terms of service requirements, everything from applications being carried, capacity, and performance are all on the sharp rise, making the consolidation and expansion capabilities of the virtualization platform all the more important. The common services carried on virtualized servers include Microsoft services, databases, Web applications, and OA. During peak usage hours, the loads on applications range from 2000 to 5000 IOPS/TB while that during moderate usage ranges from 500 to 1000 IOPS/TB. The performance density of traditional storage is 300 to 500 IOPS/TB, leaving much to be desired in terms of performance and capacity in the new environments.

2. Business continuity

A unified platform with multiple businesses has higher business continuity requirements than those of independent apps. Therefore, service continuity must be supported in virtualization scenarios. For VMware, its DR solution provides the SRM stretched active-active design.

3. Development and testing

The development and test system of virtualization platforms needs to quickly copy and access production data without adverse impact to production services. Use of LUN cloning in traditional storage is time-intensive, taking about 30 minutes to copy 10 TB of data and lowering processing capabilities in production services by 30% or more.

4. Cost reduction

Each VM in a virtualized environment takes up system resources. Resource utilization in storage needs to be improved and reduction in redundant data in the system helps lower TCO.

The Huawei server virtualization solution delivers both high performance and top levels of service continuity while reducing TCO for the enterprise.

Huawei server virtualization solution

Huawei OceanStor Dorado V3 all-flash storage provides powerful performance, supports large-scale service integration and ultra-low latency, and handles high performance density in server virtualization scenarios and the business pressure in multi-business scenarios. At the same time, Huawei solution supports active-active design with VMware SRM stretched storage to build an end-to-end high-availability virtualization platform.

Based on the thin provisioning and inline data reduction technologies, OceanStor Dorado V3 reduces costs of virtualization infrastructure platform, saves equipment room footprint and power consumption, and lowers TCO.

Highlights

1. Three times better business integration capability, with VMs bearing mission-critical business

OceanStor Dorao V3 improves the performance density by three times compared to traditional storage, and by twice compared to XtremIO X2. OceanStor Dorado V3 delivers better performance in heavily-loaded virtualization scenarios and supports 500 IOPS/100 GB for a single VM.

2. Support for VMware SRM stretched storage active-active layouts

Service continuity: Huawei gateway-free active-active solutions have acquired certification from VMware SRM Stretched Storage.

3. Three times higher development and testing efficiency based on lossless snapshots

Traditional development & test: three complete copies of data, slow speed, and high space occupation rate

Dorado V3 development & test: Lossless snapshot (no need for copying) occupies small volume of space and executes operations with zero impact to production services.

SAP HANA TDI solution

SAP HANA is a high-performance in-memory database platform launched by SAP SE. It aims to help enterprises analyze and process large amounts of data in real time and seize business opportunities. According to Gartner, a well-known third-party advisory firm, more than 100,000 SAP customers will be upgraded from traditional databases to SAP HANA databases by 2020, accounting for more than 35% of SAP users.

Although SAP HANA accelerates large-scale service innovation, SAP HANA has certain requirements for the legacy data center architecture.

1. Data center architecture

SAP HANA mainly employs the soiled construction model, bringing business island, low resource utilization, and complex business expansion. This is contrary to the trend of migrating businesses to the cloud for mainstream data centers.

2. DR and backup building

In SAP HANA appliance mode, the DR of SAP applications and database cannot be deployed in a unified manner. The database adopts software-based DR while the application adopts virtualization- and storage-based DR. The management, maintenance and DR operations are complex, which cannot ensure data consistency and cannot meet the requirements in large-scale data center scenarios.

3. Business maintenance process

Traditional databases use advanced functions of storage systems to provide data services. The SAP HANA platform supports internal enterprise businesses, involving dozens and hundreds of sub-service systems which demand for extensive development, testing, rollout, and positioning. Business innovation requires rapid release of development and testing, and sandbox systems. However, the SAP HANA system without snapshot function takes a long time to have required full copy testing for each new business rollout and upgrade. Each operation needs to be coordinated by multiple layers, such as application and hardware layers. Many operation steps make it difficult for hardware administrators to handle.

The preceding challenges are especially prominent in cloud-based data centers and in the evolution from traditional databases to the SAP HANA, making it a major challenge for enterprises to use SAP HANA system.

Huawei’s SAP HANA TDI solution

With the maturity of SAP HANA, hardware faces fewer restrictions than before. To better integrate SAP HANA into existing data centers of enterprises, SAP launches Tailored Datacenter Integration (TDI) reference architecture to integrate SAP HANA database more flexibly into existing data centers.

Huawei’s SAP HANA TDI is a solution designed based on Huawei’s powerful hardware capabilities and integration capabilities of the TDI solution. Compared to other vendors’ solutions, it has the following three features:

Scalable: SAP businesses require high hardware performance. The higher the performance is, the more HANA nodes are carried. In addition, the preceding performance data takes effect only after being certified by SAP. Huawei’s outstanding performance in all-flash storage has been recognized by SAP. In the latest SAP certification, Huawei’s high-end all-flash storage system supports a maximum of 200 SAP HANA nodes, ranking top in the industry.

Reliable: Huawei all-flash storage can still meet SAP’s high performance (1 ms latency) requirements after enabling the active-active design. Therefore, Huawei’s SAP HANA TDI solution can provide end-to-end SAP HA solutions that satisfy SAP performance requirements.

Efficient: With the lossless snapshot function, Huawei’s SAP HANA TDI solution can quickly set up the development, testing, and sandbox systems. Response time to the business environment setup is reduced from 10 hours to 20 minutes compared to traditional solution.

Huawei is the first storage vendor that puts NVMe into commercial use. Designed for enterprises’ heavily-loaded critical applications, Huawei’s all-flash storage meets storage requirements of enterprise-class applications such as databases, virtual desktops, and server virtualization, perfectly matching solid-reliability and high-performance standards in SAP HANA scenarios to build an industry-leading SAP HANA TDI solution.

The post Huawei OceanStor Dorado V3 All-Flash Solutions, Built for Your Application Scenarios appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog
—