您当前的位置是:  首页 > technology > Corporate communications > technology > Unified Communications and Collaboration > Technical dynamics >
  Home >technology > Corporate communications > technology > Unified Communications and Collaboration > Technical dynamics > Nutanix's new direction of hyper-convergence architecture

Nutanix's new direction of hyper-convergence architecture

-- Seeking breakthroughs from fierce competition — The transformation of Nutanix's hyper-convergence application concept

2018-05-14 14:07:13 Author: Source:CTI ForumComment:0  Click:


Nutanix, a pioneer and leader in hyper-convergence systems, is gradually transforming the original hyper-convergence system with storage as a comprehensive software-defined data center platform.
Over the past five years, we have witnessed the “hyperconverged infrastructure” product type, which has rapidly grown from a newly created product to one of mainstream corporate IT products. Under the impact of cloud services, a small number of people can continue to maintain high growth. IT infrastructure products.
In the face of a hyper-convergence market that is gradually becoming more and more intense, but competition is becoming increasingly fierce, Nutanix, who pioneered the field of hyper-convergence, is trying to expand the connotation of hyper-convergence systems and try to give hyper-convergence through several recent system revisions. The more diverse capabilities of the system have become a complete data center platform.
The Paradigm Shift of Hyper-Convergence System - Seeking a New Way out of Fierce Competition
Over the past few years, the hyper-convergence system is gradually moving towards its peak, but at the same time, a shadow appears.
On the one hand, although the overall sales volume of hyper-converged systems continues to grow at a high rate, the application concept and technology development itself has been stagnant. It is only in the original framework that the products of various manufacturers have become similar.
On the other hand, the high growth of the hyper-convergence market has also attracted more and more manufacturers. In addition to the newly-created companies, almost all first-tier manufacturers now have a hyper-convergence product line, and the number of suppliers is no less than 20 and 30. It also makes the competition in this product area increasingly fierce.
These two factors work together to make the development of hyper-convergence systems worse. There are more and more manufacturers and products involved, but the functions and features of products are becoming more and more similar. In the end, the cost competition will be inevitable. cycle.
In the face of this situation, Nutanix, a pioneer and leader in hyper-convergence systems, is trying to find new ways for hyper-converged applications by giving hyper-convergence systems new features and connotations.
Looking at the revision of Nutanix's AOS hyper-converged software platform over the past two years, it can be found that Nutanix is ​​transforming the original hyper-convergence system with storage as a comprehensive software-defined data center.
Traditional super fusion with storage as the core
Compared to the traditional IT infrastructure consisting of individual servers and network equipment and storage devices, the hyper-convergence innovation consists in providing node infrastructure that combines computing and storage resources to provide IT infrastructure services. In the enterprise IT environment, the same elastic configuration as the cloud environment is achieved, and the scalability of the Web-Scale level is achieved.
At the product level, the hyper-converged infrastructure is a combination of these concepts - "software-defined storage," "decentralized storage clustering," and "server virtualization platform." Standard product types are "computing. The "Integrated, Storage-In-One" Infrastructure Suite chassis is essentially a standard x86 server that relies entirely on storage clustering and hypervisor virtualization platform software to leverage and manage server nodes' computing and storage resources, and to stack more This standardized node chassis forms a cluster to obtain the necessary high availability and scalability.
Among the three key components that make up the hyper-convergence system—cluster storage system, Hypervisor platform, and x86 server, the most critical is the cluster storage system. The hypervisor platform used in hyper-convergence systems is nothing but direct use of VMware vSphere, Microsoft Hyper-V and KVM, and several mainstream platforms; in terms of x86 servers, vendors with their own server product lines such as Lenovo, Dell, Cloud Dadeng and others use their own servers directly, while the rest use white-label servers provided by third parties. So the real core technology of the hyper-convergence system is also the key to distinguishing different products. It is a storage cluster software platform.
However, after several years of development, the storage clustering functions of hyper-converged products have become similar. The storage platforms of leading vendors such as Nutanix and VMware have lost their original functions and architectural advantages.
Although some vendors have recently attempted to provide hyper-converged products with different standard types, such as NetApp HCI's computing and storage node separation architecture, or HPE SimpliVity's hardware acceleration architecture, they are still on storage.
A new generation of hyper-convergence towards a complete software definition center
Nutanix has different ideas. When other vendors are catching up to Nutanix's original hyper-converged architecture and perfecting the storage clustering platform, Nutanix is ​​ready to move toward the new hyper-converged product form.
Concerning the situation that the storage clustering function of hyper-convergence systems has matured and is not likely to develop significantly, Nutanix changed the hyper-convergence system to cover a wider range of application levels and integrate the original IT infrastructure resource management functions. Cloud management and software-defined networking capabilities allow hyper-convergence to have new meanings.
Nutanix first added a self-service automated resource management function to its AOS software platform at the beginning of 2017. Together with the cross-hypervisor management capabilities of the dedicated Prism management platform, the function of the cloud management platform was brought into the hyper-convergence system.
Then in the revision of early 2018, it also introduced a preliminary software defined network (SDN) function for the AOS software platform, allowing its hyper-converged system to reach the field of network management.
Neither cloud management nor SDN is a new feature, and related products already exist. However, Nutanix's innovation lies in integrating these functions into its hyper-converged platform, making the hyper-converged system an infrastructure that spans computing and storage. A complete software-defined data center platform for resource management, cloud management, and network management.
[Combination of Storage/Management Platform and Hypervisor] Nutanix Hyper-Convergence Platform Software Basic Architecture
[Nutanix software platform architecture] is divided into Prism, which plays the role of management platform, and Acropolis OS (AOS), which stores platform roles. It not only provides storage clustering services, but also penetrates the Hypervisor layer to provide VM management and protection functions. (Image source/Nutanix)x
In Nutanix's hyper-converged system software platform architecture, it can be divided into Acropolis OS (AOS) and Prsim. AOS is a data storage platform (or Data Plane) and is responsible for providing data storage management and application management. Prism is a management platform (or Control Plane) and is responsible for providing user interface, policy, analysis, and other services.
Which AOS is divided into 3 parts:
  (1) Decentralized storage architecture:Based on Nutanix's distributed file system (NDFS), it provides clustered storage services with advanced features such as high availability, Scale-Out scalability, and a series of I/O acceleration and data protection.
The high-availability architecture supports three-replica mirror replication across nodes and two architectures such as Erasure Coding. The expansion capability supports cluster configurations of a minimum of three nodes. The maximum cluster size is unlimited, and if availability requirements are reduced, Provides configuration for 1 or 2 nodes; advanced data service capabilities include tiered storage for I/O acceleration, snapshots for data protection applications, Clone and remote replication, and compression and deduplication to help reduce storage space.
  (2) Application mobility architecture:Penetrate the hypervisor layer to provide dynamic migration and conversion of objects such as VMs, virtual storage devices, or Containers between hypervisors, cloud environments, or Nutanix nodes to support applications such as load balancing and disaster recovery.
(3) Hypervisor: Supports the two mainstream Hypervisor platforms, vSphere and Hyper-V. It also includes Nutexix's proprietary Hypervisor, which is based on CentOS KVM and developed on its own.
Prism can also be divided into these two major parts:
  (1) User operation interface:Provides user interface for user interaction, including HTML 5 based graphical interface, REST API, CLI command line interface, PowerShell command interface, etc.
  (2) Management Services:Provides status monitoring, policy setting and analysis.
With the combination of AOS and Prism, the following four major orientation management functions are available:
  1. Provides system management of Nutanix NX series hyper-converged server hardware.
  2. Provides management of the Nutanix cluster environment, including resource configuration of the hypervisor, and automated cluster node expansion.
  3. Provides resource monitoring and management that penetrates the Hypervisor platform and goes deep into the VM hierarchy.
  4. Based on the NDFS file system, a series of data service and data protection functions corresponding to the VM level are provided, including local cross-node mirroring, snapshots and Clone, and remote site backups.
[Disclaimer] This article only represents the author's own opinion and has nothing to do with the CTI Forum. The CTI Forum maintains its neutrality in the presentation of statements and opinions, and does not provide any express or implied warranty for the accuracy, reliability or completeness of the contents contained therein. Readers are for reference only, and please bear full responsibility for yourself.

Topics