In the AI and container era, SmartX announces hyperconverged software HCI 6.0 with key upgrades, including 5 added features and 4 new scenarios, to help our customers address challenges in supporting different architecture and various new applications, managing and maintaining IT infrastructure, and replacing VMware products.
These new capabilities allow SmartX to enhance the Enterprise Cloud Platform (ECP) to help certain customers with needs to build a lightweight and open cloud infrastructure effortlessly.
Key Takeaway
New features include:
- Enterprise-grade SMTX File Storage with multi-protocol support.
- Persistent cache technology ensures mixed workload performance.
- Erasure coding and high-capacity nodes enhance storage efficiency.
- Enhanced container management capabilities to better support AI workloads.
- Upgrade center makes operations and maintenance smooth.
New use cases include:
- Distributed Storage and Application Integration Platform
- Distributed Disaster Recovery Integration Platform
- VM-Container Converged Infrastructure (VCCI)
- Lightweight AI Infrastructure
What’s New
Multi-Protocol Support: Enterprise File Services
Introducing SMTX File Storage, offering enterprise-grade file storage for hyperconverged infrastructure. Users can store and manage unstructured data like text, images, and videos simply and efficiently without the need for separate NAS storage, reducing additional equipment costs and management overhead.
SMTX File Storage leverages SmartX’s years of technology in storage, offering the following advantages:
- Flexible and Reliable: Inherits stable core of hyper-convergence, high availability of file controllers, and scalability at node or file controller level.
- Flexible and Open: Compatible with major operating systems, supports Kubernetes NFS CSI Driver, and is accessible both internally and externally within clusters.
- Easy Operations: Quick deployment in 10 minutes, one-click deployment, online upgrade and scaling, visual service status and anomaly alerts.
SMTX File Storage can be used with SMTX OS Standard and Enterprise editions, and included in SmartX ECP Standard and Advanced editions, at no extra licensing cost.
Mixed Workload Performance Assurance: Persistent Cache Technology
SmartX utilizes a single storage resource pool to provide block and file storage services simultaneously. By introducing persistent cache technology, it significantly improves performance for both block and file data, delivering performance comparable to all-flash arrays.
- Block Storage: Virtual machines running performance-sensitive applications such as critical business and database systems can keep data resident in the high-speed cache layer, ensuring stable high performance and safeguarding essential business operations.
- File Storage: Applications mounting file systems can keep file system metadata resident in the cache, improving file system access performance, reducing access latency, accelerating data access speed, and ensuring application performance for mounted file systems.
- Other Applications: For other storage-dependent operations, the system dynamically adjusts tiering strategies based on cache usage and IO access using efficient hot and cold automatic tiering technology, continuously migrating cold data to capacity tiers to ensure efficient use of cache resources.
After storage performance optimization, SmartX can deliver significant read and write performance when providing block and file storage services to users.
Storage Efficiency Enhancement: Erasure Coding and High Capacity Nodes
Through features like erasure coding, SmartX helps users reduce overall storage costs while balancing data value and storage expenses.
Erasure coding supports 2+1, 4+1, 4+2, 8+2, and 28 other configurations, allowing flexible selection of EC ratios based on cluster size. Compared to replication mechanisms with the same redundancy capability, erasure coding significantly improves space utilization (theoretical minimum of 66%, up to 96%).
Single-node capacity supports up to 256 TiB, reducing average storage costs per unit in clusters and efficiently supporting large-capacity storage scenarios. It also supports reclaiming storage space from unused thin provisioned volumes, enhancing storage resource utilization.
Enhanced Container Management Capabilities: AI Compute Support
On the foundation of unified management of virtualization and containers, SmartX introduces the following capabilities to assist users in deploying container applications on hyper-converged systems, addressing fluctuating AI compute demands and potential node-level challenges:
- Automatic allocation and mounting of GPU/vGPU: Automatically mounts suitable GPU/vGPU devices for Kubernetes clusters.
- Elastic scaling and automatic replacement of nodes: Automatically scales GPU/vGPU resources up or down and supports automatic replacement of failed nodes.
- GPU partitioning and sharing: Enhances GPU resource utilization through technologies like Time-Slicing, MIG (Multi-Instance GPU), and MPS (Multi-Process Service).
Operations Capability Upgrade: Upgrade Center
The Upgrade Center provides independent pre-upgrade checks to identify potential upgrade risks in advance. The upgrade process is more automated, guiding users through a simple interface and one-click operations to manage upgrades for kernel, operating system, and patches.
Supporting New Use Cases
Distributed Storage and Application Integration Platform
Platforms for storing user images, audio, and video data often require different types of storage resource pools (hot, warm, cold) to balance performance and capacity. They also need computing resources for data collection, application, and intelligent analysis.
SmartX introduces an integrated solution that provides high-performance and balanced storage within a single system, eliminating the need for separate compute nodes and reducing complexity and space requirements.
For example, for heathcare PACS application,, this platform offers high-performance storage for recent medical images and balanced storage for long-term images. Applications can flexibly allocate resources, leveraging GPU and container management for intelligent analysis. This simpler architecture ensures efficient data application, lower costs, and higher operational efficiency, maximizing IT cost-effectiveness.
Distributed Disaster Recovery Integration Platform
The SmartX Distributed Disaster Recovery Integration Platform offers tailored, cost-effective backup solutions with higher performance and scalability, reducing overall TCO. It integrates with virtualization technologies for a two-city, three-center disaster recovery setup.
For medium and large customers, SmartX provides a distributed backup resource pool without separate backup hardware. Backup services can run directly on high-capacity hyper-converged nodes, supporting various applications. Each node offers over 1GiB write performance and more than 2TiB per hour recovery speed.
For small and medium-sized customers, SmartX partners with CDP vendors to offer a distributed disaster recovery integration solution. This solution provides data protection and second-level RPO for SmartX VMs, VMware VMs, and physical machines, along with the necessary compute and storage resources for backup and emergency recovery.
Additionally, SmartX virtualization network, based on Overlay technology, seamlessly extends business networks across different clusters, racks, data rooms, and data centers. This enables rapid VM network recovery across multiple sites and clusters, reducing RTO and minimizing business interruption, ensuring business continuity.
VM-Container Converged Infrastructure (VCCI)
To better support mixed virtualization and container workloads, SmartX introduces the VM-Container Converged Infrastructure (VCCI) solution based on its full-stack hyper-converged capabilities. This unified architecture provides server virtualization, distributed storage, networking and security, and container management, helping users efficiently deliver various foundational resources (general compute, AI compute, storage, network). It enhances resource utilization and simplifies VM and container network interconnectivity.
The VCCI solution is already deployed in production environments across multiple industries, including healthcare, manufacturing, and transportation. It allows enterprises to support both virtualization and container application architectures with higher resource efficiency, agile deployment, and enhanced security for applications, data, and networks, helping businesses tackle the challenges of digital transformation.
Lightweight AI Infrastructure
AI applications in enterprises demand diverse and comprehensive support for CPU and GPU compute power, storage capacity, and performance. SmartX introduces a lightweight AI infrastructure with three key features: compute convergence, workload convergence, and storage convergence, effectively addressing these challenges:
- Compute Convergence: Flexibly configure hosts with different numbers and models of GPUs. SmartX supports GPU passthrough, vGPU, MIG, and MPS, allowing flexible GPU partitioning to enhance utilization and AI application efficiency.
- Workload Convergence: With SMTX Kubernetes services, support both virtualized and containerized AI applications. Integrated deployment optimizes resource allocation, meeting diverse requirements for performance, security, scalability, and agility, improving resource utilization and reducing costs.
- Storage Convergence: Provide both block and file storage for AI applications. Features like IO localization, data tiering, Boost mode, RDMA support, and persistent cache ensure high storage performance. SmartX’s storage technology supports various hardware options and flexible scaling, balancing business needs and costs effectively.
SmartX’s customer Connectwave shared their story of deploying SmartX to support AI applications.
About SmartX
SmartX is recommended in Gartner’s “Market Guide for Full-Stack Hyperconverged Infrastructure Software” and the Asia Pacific customers’ choice in “Voice of the Customer for Full Stack Hyperconverged Infrastructure Software”. With over 1000 customers and over 20000 deployed nodes, SmartX holds the No.1 place in HCI software market share in China. SmartX has deep collaboration with customers including Bank of Communications, Guotai Junan Securities, Foxconn, Cafe 24, etc., with one customer’s install base exceeding 1000 nodes.
