HUAWEI OceanStor N8500 V200R001C10 Main Slide
Agenda 1 Trend & Challenges 2 Specifications & Features 3 Application Scenarios 4 Stories 2
Global Data Trend 2020: 2009: 50X 0.8 ZB 40 ZB Source: IDC 2012 The global data volume was 0.8 ZB in 2009 and will be up to 40 ZB in 2020. What does 40 ZB mean? 40 ZB ≈ 2 trillion Blu-ray Discs (BDs) (20 GB for a single BD) 40 ZB ≈ 5247 GB data for each person (about 7.6 billion population in 2020) 3
Challenges Faced by Enterprise Customers Excessive services Fast data growth Low storage efficiency Heavy pressure on Difficult storage Large performance expansion investment 4
Agenda 1 Trend & Challenges 2 Specifications & Features 3 Application Scenarios 4 Stories 5
HUAWEI OceanStor N8500 Overview • Excellent performance, 3 million OPS in the SPEC benchmark test • Up to 15 PB flexible storage space and PB- level file storage, fitting for high-speed service growth • Efficient data management, decreasing the TCO by 60% • Intelligent and converged storage, simplifying the infrastructure 6
OceanStor N8500 Specifications Item N8500 System Multi-node active-active cluster architecture architecture Cabinet Number of nodes 2 to 24 (optional) Maximum system GE switch 15 PB capacity (optional) Basic: 16 GB; Standard: 24 GB Fibre Channel Cache per node Enterprise: 48 GB; Enhanced: 96 GB (standard configuration) and switch (optional) scalable to 192 GB Ethernet GE and 10GE Engine Protocol NFS, CIFS, FTP, HTTP, FCP, FCoE, iSCSI, NDMP LAN-Free Storage unit Maximum capacity 256 TB of a file system DST, LoadBalance, Snapshot, Mirror, Quota, Replication, WORM, Value-added SmartCache, HyperThin, HyperImage, HyperCopy, HyperClone, and software HyperMirror Disk type SSD, SAS, NL-SAS, SATA, and intermixing of different disks 7
OceanStor N8500 Architecture Node 2 Node 1 Node 3 Node N Service cache Service cache Service cache Service cache NAS 10GE GE 10GE GE 10GE GE 10GE GE Control Service cache Service cache Service cache Service cache engine plane FC FC FC FC CPU CPU CPU CPU CPU CPU CPU CPU Data plane 10GE GE 10GE GE FC 10GE GE FC FC 10GE GE FC Service cache Service cache Service cache Service cache Service cache Service cache Service cache Service cache Storage Service cache Service cache Service cache Service cache Service cache Service cache Service cache Service cache unit CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU CPU JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD JBOD Separation of the control plane and data plane Separation of file management and disk management Fully redundant architecture Fully modular design 8
SPECsfs2008 nfs: 3,064,602 OPS Apr. 2011 Oct. 2009 Oct. 2012 The Standard Performance Evaluation Corporation (SPEC) is a global authoritative third-party organization that tests application performance. Operation per second (OPS) refers to the number of supported concurrent requests per second. 9
Excellent performance — Fully redundant clustered architecture Traditional NAS architecture: non-cluster, Clustered NAD architecture: active-active, non-redundant, limited concurrent fully redundant, strong concurrent processing capability processing capability a a b c d b Server … … … … Network NAS engine A NAS engine B NAS engine A NAS engine B NAS engine C NAS engine D File system A Storage File system A Storage pool File system B File system B 10
Cluster Parallel processing, improve the performance The standard file CIFS CIFS NFS NFS CIFS NFS access protocol, wide compatibility CFS CFS CFS The global lock Cluster Cluster management, to GLM GLM GLM heartbeats heartbeats ensure the data CVM CVM CVM consistency Private communication GAB GAB GAB protocol, to ensure that the information LLT LLT LLT transmission efficiency and consistency 11
Excellent performance — I/O paths with the same length and latency Clustered NAS architecture: data sharing, I/O Traditional NAS architecture: data forwarding, I/O paths with the same length and latency, stable paths with different lengths and latencies performance a c d a b c d b Server Network HA HA NAS engine B NAS engine C NAS engine D NAS engine A NAS engine A NAS engine C NAS engine B NAS engine D Storage Virtual storage resource pool A A B B C D D C A B A B D D C C 12
High scalability — Flexible and Cost-effective Expansion Modes Traditional mode: limited, restricted Modular mode: flexible, cost-effective … … … … … … Server … … … … … … Network Storage NAS engine Max. 24 nodes File system A File system A Storage unit File system A File system A File system A Max. 15 PB 13
Robust reliability — Max. N – 1 node redundancy Clients IP resource pool 192.168.1.10 192.168.1.10 192.168.1.12 192.168.1.11 192.168.1.10 192.168.1.11 192.168.1.10 192.168.1.12 192.168.1.11 192.168.1.11 192.168.1.11 192.168.1.10 192.168.1.13 192.168.1.13 192.168.1.13 192.168.1.12 192.168.1.12 192.168.1.13 192.168.1.12 192.168.1.13 Controller A Controller B Controller C Controller D Controller B Controller B Controller C Controller C Controller D Controller D Controller A Controller D Controller A Controller B Controller C Controller A Virtual storage resource pool 14
Load balancing Balance policy: round robin – based, CPU utilization, N8500.com Built-in DNS N8500 Balance N8500 domain domain name node IP policy network bandwidth, number of client links name resolution address 40 clients 192.168.1.10 Virtual IP 192.168.1.11 resource 192.168.1.10 192.168.1.12 pool 192.168.1.11 192.168.1.12 192.168.1.10 192.168.1.13 192.168.1.11 192.168.1.13 192.168.1.12 192.168.1.13 20 clients 10 clients 10 clients 10 clients 10 clients 10 clients 10 clients NAS engine C NAS engine D NAS engine D NAS engine B NAS engine B NAS engine C NAS engine A NAS engine A Virtual storage resource pool 15
Load balancing 120 clients 192.168.1.10 Virtual IP 192.168.1.11 192.168.1.12 192.168.1.11 192.168.1.10 192.168.1.19 192.168.1.10 192.168.1.13 192.168.1.16 resource 192.168.1.16 192.168.1.19 192.168.1.12 192.168.1.13 192.168.1.20 192.168.1.11 192.168.1.14 192.168.1.17 pool 192.168.1.17 … 192.168.1.20 192.168.1.14 192.168.1.21 192.168.1.12 192.168.1.15 192.168.1.18 192.168.1.18 192.168.1.21 192.168.1.21 192.168.1.15 40 clients 30 clients 30 clients 40 clients 30 clients 40 clients 30 clients NAS engine A NAS engine B NAS engine B NAS engine A NAS engine C NAS engine D NAS engine C NAS engine D Virtual storage resource pool 16
Intelligent migration, improving performance while reducing costs DST: SmartCache: Smart relocation of non-hotspot data reducing costs by 60% Accelerating hotspot data access by 10 times … … Tier 1 SAS Tier 2 SATA Server Network Storage Hotspot data Cold data Tier 0 SSD Tier 1 SAS 17
Multiple Features Ensuring Data Security Remote replication Incremental data – based File system asynchronous replication mirroring Fibre Channel or IP link Local HA solution Replication link Snapshot Device-level compression redundancy Preventing deletion Administrators by mistake Service continuity Snapshot data used protection Master for test and System verification Storage 18
Local High availability Local High availability & & Async Async Replication Replication Site A Site B 1. Local high availability .The business will not interrupt if one storage failed. 2. Async data replication, to ensure the data security if site A failed. WAN Engine B Engine D Engine B Engine A Engine C Engine A File System Mirror Async Replication (FC/IP) Mirror A Mirror B Storage Pool Storage Pool 19
Multi-protocol Convergence —— Reducing Investment Costs Management File-based Block-based Others interface NFS FC SMI-S NDMP CIFS SNMP NTP iSCSI FTP FCoE Syslog SSH HTTP OceanStor N8500 storage system 20
Agenda 1 Trend & Challenges 2 Specifications & Features 3 Application Scenarios 4 Stories 21
Application Scenario — high performance file processing for M&E Editing & Production Sampling NSP Special Editing Composition Catalog Inde effects Core Node x Archive Delivery Index HD: 100 Mbit/s SD:50 Mbit/s MAM N8500 N8500 Broadcast service Process Web Transcoding Edge Node Edge Node Edge Node Tapes … Strea Strea ming ming Broadcast N8500 system Typical cases Requirements Advantages Editing & Production: low latency, high reliability, Large bandwidth and high cost performance: proven 20% to Media asset storage system: scalable from 200 TB to several PB, 7Flexible expansion ， cost-effective scale up and scale out easy management configuration Flexible expansion of both performance and capacity 0% higher performance than other vendors Verified seamless compatibility with industrial ISV Verified compatibility with industrial ISVs such as AVID, Sobey, Jetsen, and DAYANG 22
Recommend
More recommend
