Cloud Storage solution

In terms of user business, the cloud may face many challenges and pain points in data storage, such as designing a virtualization architecture for non-cloud storage or a proprietary hardware architecture, which cannot be maximized when paired with a cloud platform. The scenario is single, the cost is high, the scalability is poor, upgrading is difficult, operation and maintenance are complex, and it is easy to become a data island. The EasyStack Cloud Storage solution is a storage form based on the concept of cloud computing, designed for users to better use storage in the cloud. The acquisition, activation, and use of storage capabilities exist in a service-oriented manner, eliminating the dependence of data storage on traditional hardware. At the architectural level, cloud storage is built on a unified cloud computing architecture, fully compatible with the open-source ecosystem, and has production-level reliability and product-based user experience. It supports different hardware architectures including mixed flash, full flash, and mechanical; Supports different architectures such as x86/c86/ARM, supports iSCSI and rbd storage protocols, adapts to diverse computing power and cloud-native scenarios, and is more suitable for building large-scale cloud computing infrastructure. At the cost level, cloud storage fully abstracts and encapsulates the infrastructure through the EOS digital native engine. By fully integrating key technologies such as software-defined computing power, software-defined networks, software-defined storage, and cloud-native, it brings the maximum release of comprehensive performance and the lowest total cost of ownership in heterogeneous hardware architecture, diverse storage usage scenarios, and different application scenarios. At the performance level, cloud storage provides higher performance and lower latency storage services by fully optimizing software stacks, IO paths, network transmission links, data structures and algorithms, fully leveraging high-performance hardware capabilities, and utilizing flexible product architecture combinations.

Architectural advantages

Software definition

Software definition, hardware decoupling, and continuously enriching product capabilities based on the open-source ecosystem.

Efficient metadata management

Centralized management of cache metadata and distributed management of capacity storage metadata.

High scalability

Storage node 3-node starting configuration, unlimited support for multi-cluster expansion.

Excellent efficiency

Multi-level cache architecture, industry-leading performance. High-performance and low-latency block storage services can be combined with capacity-based block storage to form a layered storage architecture. Capacity-based block storage services are built on an open-source ecosystem, providing distributed block storage with large capacity, high reliability, and high scalability for cloud platforms.

Distributed Access Services

The IO bottleneck is no longer affected by the access layer and can be horizontally expanded to meet the IO access needs of medium to large-scale multi-computing power.

Sustainable lifecycle

One-time procurement, flexible expansion, capability upgrade, and evolutionary capability.

Functional characteristics

Evolvable cloud storage architecture

Cloud-native architecture, distributed, storage cloud service-oriented, multi-computing power support, and software-defined smooth upgrades.

Integrated management, operation and maintenance

Hardware ready, minute level service activated on demand; Unified monitoring, logging, and alarm services; Unified permissions, console service configuration, and access.

Flexible deployment form

Support the architecture of integrating storage and computing (capacity-based) and separating storage and computing (performance-based+capacity-based); Supports one cloud multiple pools, and multiple cores; Supports mainstream hardware specifications for mixed flash and full flash.

Financial grade stability, reliability, and high availability

Multi replica node level protection and automatic data recovery; Rack level reliability deployment plan; Intelligent data recovery strategy with multi-level business priority; Real-time detection of hard disk anomalies and automatic isolation of slow disks.

Excellent performance

Simplify data models, deeply optimize IO paths, achieve low latency in 100 microseconds, and achieve 300000 IOPS per node; Separation of hot and cold data, intelligent data backflushing, QoS intelligent configuration for replica recovery, reducing read and write amplification; Small IO aggregation, controllable large IO passthrough, optimizing cache throughput; Full cache automatic garbage collection.

On-demand expansion

Starting from 3 nodes; Performance and capacity storage can be separated and expanded as needed; Linear growth in performance and capacity.

Safe and reliable

Distributed Multiple Copy Guarantee Mechanism

Three copies of data are highly consistent and reliable, with fully symmetric distribution of data copies to ensure data security.

Disk verification and slow disk isolation

The storage engine verifies the consistency of device data and identifies bad silent data. Simultaneously perceive slow IO and automatically isolate.

Multi-level configurable business recovery

The data recovery strategy for fault scenarios and expansion scenarios can be set through multiple levels to ensure a balance between business and recovery. The bandwidth for fault recovery scenarios can be configured.

Decentralized Access Layer Controller

The access layer controller is decentralized, without access gateway bottlenecks and single point of failure risks.

Node-level protection and intelligent data reconstruction

Distributed data protection with multiple replicas between nodes, preventing the spread of fault scenarios. Fault scenario data can be incrementally restored.

No perception processing for maintenance scenarios

Planned proactive operation and maintenance scenarios such as restart, shutdown, and disk replacement, with no sense of business and processing in seconds.

Multiple IO path access

By combining switch stack redundancy and network card bonding technology, high availability of access paths is provided through multi-path access via iSCSI.

Data degradation protection mechanism

Data degradation timeout protection, automatically restoring data replica status, ensuring data security. At the same time, the global configurable timeout period balances business continuity and data security requirements.

Cluster performance balancing strategy

Can be configured with a cluster bandwidth performance allocation strategy to ensure the acceleration of performance-based storage cluster throughput, while reducing peak business cluster impact.

Cloud products

EasyStack provides a unified construction solution for cloud integration, and cloud storage can efficiently support user business by combining computing, network, and other cloud products. Currently, multiple cloud products have been released for computing, storage, network, cloud-native, monitoring and operation, and cloud resource management. Users can flexibly choose according to compliance, management needs, application needs, technical architecture, etc., so that they can truly achieve on-demand construction, trial, and expansion of cloud infrastructure.

---

High performance computing application data storage

Background requirements

High performance computing tasks, high concurrency processing, require support for high-performance storage to provide high bandwidth, high IOPS response capability, and low latency.

Scheme capability

EasyStack Cloud Storage solution has high-performance features, with a single volume capacity of over 80000 IOPS and a single volume throughput of over 1GB/s; Having low latency characteristics, the IO access latency can be below 1ms; At the same time, it also has the feature of large capacity, with a single volume capacity of up to 64TB.

Database application data storage

Background requirements

Database applications belong to read-write intensive application scenarios, and various types of databases and data warehouses such as SQL Server, MySQL, SAP HANA, etc. require data storage to provide high-speed and low latency responses for small IO and random IO.

Scheme capability

EasyStack Cloud Storage solution, integrating storage and cloud construction, optimizes software stack and IO path, streamlines data models, provides ultra-high IOPS capability and ultra-low latency response.

Persistent storage of cloud-native application data

Background requirements

User cloud-native applications meet the needs of sensitive businesses, and for stateful application data, persistent storage is required.

Scheme capability

EasyStack Cloud Storage solution provides highly compatible persistent storage services for cloud-native application scenarios. Cloud-native is ready, and cloud storage can be seamlessly used without the need for integration with the EasyStack cloud-native platform application; CSI support, CSI interface enables cloud-native application data access, and provides CSI direct volume method for Kata secure container applications to improve access efficiency; Unified integration, unified integration with the native platform of EasyStack Cloud, unified base, unified management, operation and maintenance.

User core cluster application data storage

Background requirements

The user core cluster application has high requirements for data storage, and requires high-performance, shareable, and highly reliable storage services while ensuring data security.

Scheme capability

EasyStack Cloud Storage solution provides reliable data storage for such application scenarios. Equipped with high-performance features, combined with shared cloud hard drives, providing higher performance for cluster applications; Having the feature of multi-instance sharing, supporting mounting to multiple computing instances in a shared manner, and supporting cluster applications; Featuring high reliability, 3 copies of data, strong consistency, automatic self-healing in case of control surface failure, and automatic data reconstruction in case of data surface failure.

Enterprise application system data storage

Background requirements

Enterprise application systems, such as SAP, OA, ERP, etc., meet the diverse office needs of enterprise organizations. The data storage used should be in line with the current practical needs of users and take into account future business development.

Scheme capability

EasyStack Cloud Storage solution, with high reliability, 3 copies of data, automatic data reconstruction and self-healing to ensure storage continuity; Can meet high scalability needs, cloud hard drives support online scalability, storage nodes can be expanded online, and performance can grow linearly; Large capacity, with a single volume capacity of up to 64TB; Uniformity, the same storage pool can provide capacity or performance storage space as needed, and the space ratio can be adjusted through software to meet the dynamic development needs of diversified businesses.

Developing and testing business data storage

Background requirements

Data storage services required for developing and testing business, support user deployment and development of testing applications, and meet the needs of development, testing, deployment, and verification.

Scheme capability

EasyStack Cloud Storage solution has the advantages of high reliability, high scalability, and large capacity, while also considering cost. The required capacity storage space can be located in the same storage pool as the performance storage space required for other production applications, without the need for separate construction.