Recently, IDC released the "China Software-Defined Storage (SDS) and Hyperconverged Storage Systems (HCI) Market Quarterly Tracking Report , First Quarter of 2023" (hereinafter referred to as the "Tracking Report"). Different from previous reports that mainly focused on integrated hyper-converged systems, this time IDC conducted a separate market analysis on hyper-converged software integrated systems (HCI Software Certified System, that is, integrated systems driven by independently sold hyper-converged software). Similarly, Gartner, which has focused its Magic Quadrant report on the hyper-converged market on "hyper-converged software" since 2020, also released the "Full-stack Hyper-converged Software Market Guide" (hereinafter referred to as the "Market Guide") in March this year. "Software" also clarifies the development trend of "full stack" .
Why are the two authoritative organizations emphasizing hyper-converged software capabilities and delivery forms? What capabilities does “full stack” hyper-converged software impose on manufacturers? This article will combine the contents of the report for in-depth analysis.
Return to the essence: hyper-converged infrastructure is a product form of software-defined data center
Regarding "hyper-converged infrastructure", many users often have the misunderstanding that hyper-convergence is the integration of software and hardware and must be delivered as an all-in-one machine. This is related to the mainstream sales form of hyper-converged products - traditional server manufacturers provide all-in-one machines, which is the hyper-converged integrated system (HCIS) defined by Gartner. But in fact, as a product form of software-defined data center (SDDC), hyper-convergence is essentially a "software-defined" architecture.
TechTarget pointed out in "Software-Defined Everything" that the core of "software definition" is the abstraction and virtualization of hardware resources: "'Software definition' uses virtualization technology and abstracts workloads from the underlying hardware, thereby allowing The IT infrastructure is more flexible and agile". Under this model, the deployment, operation, supply and configuration of virtual resources are all managed by software, which not only improves resource utilization and efficiency, but also reduces the IT infrastructure's dependence on hardware, thereby reducing hardware investment and accelerating technology iteration. A software-defined data center is a data center built on a software-defined architecture—all infrastructure elements, including computing, network, and storage, are virtualized and delivered as a service.
Hyper-converged infrastructure is a product form of software-defined data centers, so it also has the characteristics of "software-defined". This is evident from the definition of hyperconvergence in Gartner's Market Guide: "Hyperconvergence is a software-defined infrastructure that provides virtualized compute, storage and networking through a single instance running on server hardware". At the same time, hyper-convergence also decouples storage software from hardware through software-defined storage (SDS), eliminating the software's dependence on proprietary hardware and further leveraging the advantages of "software definition".
It can be seen that the following advantages of the hyper-converged architecture are all given by "software definition":
- Software-driven : The deployment, operation, provisioning, and configuration of virtual resources are managed by software to improve resource utilization and efficiency.
- Rapid iteration : Unbound hardware, based on the latest server hardware, software upgrades can quickly drive IT infrastructure functions and performance improvements.
- Simplified architecture and management : The data center uses a unified computing and network architecture, and O&M personnel do not need to master proprietary storage hardware or storage network O&M technology.
On this basis, the hyper-converged architecture also has the architectural advantages of fully distributed and converged deployment, which further improves the elastic expansion capability of the system, simplifies the IT architecture, and greatly reduces the total cost of ownership and the complexity of operation and maintenance.
Software delivery or appliance delivery?
It is precisely based on the characteristics of "software definition" that after long-term development, the hyper-converged market has become obviously polarized in terms of delivery forms: hyper-converged integrated systems are still dominated by hardware providers, while on the other hand, hyper-converged software providers use The feature of hyper-convergence "decoupling software and hardware" has introduced and dominated the delivery form of pure software.
Both delivery methods have pros and cons. The biggest advantage of all-in-one machine delivery is "ready-to-use" - manufacturers provide users with hyper-converged solutions integrating software and hardware based on customer needs and their own product strategies. This delivery method is simple and fast, and manufacturers and users can conduct more fine-grained management of fixed hardware platforms. Manufacturers will also provide customers with firmware-level product verification, compatibility verification, and upgrade guarantees.
However, the delivery method of all-in-one PCs also limits users’ hardware choices. This is precisely the advantage of software delivery: large customers can flexibly choose hardware suppliers based on their own hardware cost advantages. At the same time, users can purchase independent software to reuse existing hardware. If it is based on software with strong support for heterogeneous (different configurations and brands) servers, users can integrate servers of different generations and brands without worrying about hardware upgrades and supplier replacements, giving full play to the advantages of "software definition."
Image source: Getting started｜Hyper-converged model selection: pure software or all-in-one machine? ,Know almost
Gartner also emphasized in the "Market Guide" that "software delivery is more friendly to special application scenarios and special needs in different regions (such as Xinchuang), and can also provide the latest capabilities for cloud and Kubernetes-related applications." IDC believes that "industries that have high requirements for IT infrastructure flexibility, such as the financial and communications industries, should consider the form of software delivery to flexibly deploy resources while reducing TCO."
It is also worth users' attention that in the domestic market environment, most traditional IT manufacturers such as Huawei and H3C only provide the delivery method of hyper-converged all-in-one machines, and the adaptation to the hardware is not extensive enough. Independent hyper-converged vendors represented by SmartX have self-developed hyper-converged software and good hardware compatibility, and can support different delivery forms for users to choose flexibly.
Full-stack hyper-converged software: testing the key capabilities of hyper-converged product suppliers
As enterprises increasingly tend to adopt a full set of infrastructure services provided by a supplier, the "full stack" capability of hyper-converged software has also become an important criterion for Gartner to evaluate hyper-converged vendors. Gartner defines it in the "Market Guide": "Full-stack hyper-converged software provides a complete software solution, that is, in a single instance, covering virtualized computing, storage and network functions. Full-stack hyper-converged software vendors are those that can develop and a hyper-converged software vendor that sells server virtualization, software-defined storage and network management tools. ”
In addition to the above three main component support, full-stack hyper-converged software should also provide users with operation and maintenance management functions, Kubernetes support, and edge deployment capabilities:
- Intelligent operation and maintenance : with monitoring, security, management, multi-environment optimization and management functions.
- Kubernetes support : Provides container management capabilities, and manages virtualization and Kubernetes clusters through a hyper-converged management platform.
- Edge Deployment : Supports simple deployment and O&M in data centers, colocation rooms, edges or public clouds.
The complete full-stack hyper-converged software capabilities are shown in the figure below.
Image source: Market Guide for Full-Stack Hyperconverged Infrastructure Software, Gartner (2023)
Gartner put forward the following specific requirements for these capabilities in "Critical Capabilities for Hyperconverged Infrastructure Software":
- Hardware: Hyper-converged software products should be widely compatible and integrated with various hardware platforms and components.
- Virtualization : Provides extensive and in-depth virtualization technologies for hybrid cloud usage scenarios.
- Containers : Aiming at cloud-native usage scenarios, it provides extensive and in-depth container support technologies.
- Storage (data service) : It can provide high-performance storage functions and services for all usage scenarios, and provide high availability through backup, disaster recovery and other solutions. It has flexible expansion capabilities within and across nodes.
- Network and security : with network security and data compliance functions.
- Operation and maintenance management : It can provide simple, feasible and differentiated operation and maintenance management tools for all usage scenarios.
- Software technology stack : Provide self-developed full-stack software and extensively integrate with third-party software stacks to provide multiple delivery models for different applicable scenarios.
- Deployment locations : Hyper-converged solutions that provide cost-effective, high-performance, high-availability, and easy O&M for remote offices, branch offices, and edge locations.
As the first full-stack hyper-converged software market report released, Gartner also analyzed representative vendors such as Nutanix, VMware, Microsoft, SmartX, and Sangfor. Among them, SmartX ranks first in the hyper-converged independent software market in China with a market share of 21.9% in IDC's "Tracking Report" .
Overview of SmartX hyper-converged full-stack software capabilities
Based on the self-developed core distributed block storage ZBS and native virtualization ELF, SmartX provides customers with a SmartX Ready Nodes deployment solution: users can use the SmartX hyper-converged software SMTX OS and select certified servers in the compatibility list for deployment. This solution can provide users with complete hyper-converged software capabilities that meet Gartner’s “full stack” requirements:
- Hardware openness: Supports more than 20 types of servers that meet Xinchuang requirements, including Haiguang, Kunpeng, and Feiteng chips, and nearly 40 types of mainstream x86 and ARM servers. Users can also reuse existing servers or integrate newly purchased servers of different brands and configurations to reduce system construction costs. To learn about user practice cases, please read: 4 hyper-converged user practices that reveal how to achieve architecture transformation at a lower cost .
- Multiple virtualization platform support: Users can retain their original usage habits and use VMware virtualization (or Citrix virtualization in VDI scenarios), or they can choose SmartX native virtualization ELF to further save licensing fees and realize domestic replacement of virtualization. .
- Self-developed core distributed storage: does not rely on any open source distributed storage software, and continuously iterates to achieve the evolution of enterprise-level functions and end-to-end performance optimization. Its performance, stability, and reliability have been tested in the production environments of many leading customers in finance, medical, manufacturing and other industries, and can efficiently and stably support core production business systems, databases, disaster recovery, development testing, edges, containers, etc. Application scenarios.
- Network and security based on "zero trust": Micro-segmentation technology based on the zero-trust network model can provide a distributed firewall for the cluster, and use whitelist mode, security policy following, "label" and "security group" management methods, Features such as "one-click isolation" of suspicious virtual machines ensure the security of "east-west" traffic in the virtual environment. To learn more about product features, please read: How to implement east-west access control in a virtualized data center .
- Kubernetes support: Combined with the production-level Kubernetes construction and management service product SMTX Kubernetes Service (SKS for short), users can build Kubernetes clusters on hyper-converged platforms using ELF virtualization to provide high-performance and stable operating environments for stateful applications. Automated and standardized full lifecycle management of Kubernetes clusters.
- Unified operation and maintenance management of multiple environments: Combined with the SmartX hyper-converged cluster management platform CloudTower and SKS, users can achieve unified management and operation and maintenance of virtualized applications and container applications through a set of management planes, improving the efficiency of multi-environment cluster management.
- Disaster recovery and backup: SmartX provides users with hyper-converged disaster recovery solutions that can meet various levels and scenarios, including cluster high availability, backup and recovery, CDP (continuous data protection), asynchronous disaster recovery and synchronous active-active, protecting While maintaining the overall data of the cluster, it can guarantee the high availability of key businesses. For more details about the solution, please read: Learn about the SmartX hyper-converged disaster recovery backup solution .
For more SmartX hyper-converged full-stack software capabilities and features, please scan the QR code below to download "SmartX Hyper-Converged Technology Principles and Feature Analysis Collection (including VMware Comparison)" to learn more about how SmartX improves infrastructure performance and reliability through technological innovation sex.
Reference article:
1. Market Guide for Full-Stack Hyperconverged Infrastructure Software, Gartner, 2023
https://www.gartner.com/document/4178199
2. China Software-Defined Storage (SDS) and Hyperconverged Storage System (HCI) Market Quarterly Tracking Report, Q4 2022, IDC, 2023
https://www.idc.com/getdoc.jsp?containerId=prCHC50721823
3. Software-Defined Everything (SDE), TechTarget
https://www.techtarget.com/searchdatacenter/definition /software-defined-everything-SDE#:~:text=Software%2Ddefined%20everything%20(SDE)%20is%20an%20umbrella%20term%20that,infrastructures%20more%20flexible%20and%20agile
. Hyperconverged Infrastructure Software, Gartner, 2021
https://www.gartner.com/document/4008403
5. Getting Started｜Hyperconverged Infrastructure Software: Pure Software or All-in-One Machine? , Zhihu
https://zhuanlan.zhihu.com/p/128571887