Differences and relationships among hyper convergence, software defined storage (SDS), distributed storage and server San

Technical concepts that dazzle people for a while

Storage devices have always been an important part of computer systems , A long time ago , Users are used to using SAN Storage is the only choice for storing enterprise critical business data . be based on SAN A three tiered infrastructure with storage at its core （ The server - Storage networks - The storage device ）, Become the most common infrastructure in the enterprise , Also known as traditional architecture . The emergence of the Internet , The amount of information is growing explosively , Human beings are producing data faster than all previous eras combined , The amount of data that needs to be stored is also growing rapidly . big data , Artificial intelligence , New technologies such as the Internet of things have emerged , New requirements for storage . A series of new technical concepts and terms about storage have emerged in the market （ Include Software definition storage (SDS)、 Hyperfusion (HCI)、 Distributed storage 、Server SAN wait ）, It rushed into the eyes of users ; Faced with some specious definitions , Description of the functions of various manufacturers , I believe you would like to know what is the essence behind these concepts ？ Can solve those problems ？ And whether there is a suitable application scenario ？

For the above reasons , This article will introduce several concepts mentioned above , Help people better understand these technologies or concepts , Make a choice that suits you .

Conceptual explanation

Software definition storage (SDS) What is it?  

SDS The full name is Software Defined Storage , The literal translation is software definition storage . About SDS Refer to the global networked storage industry association for the definition of （Storage Networking Industry Association,SNIA）,SNIA stay 2013 Officially Software definition storage (SDS) Included in the study .

SNIA Define storage for software (SDS) Is defined as ： A virtualized storage with a service management interface . SDS Including storage pooling , The data service characteristics of the storage pool can be defined through the service management interface . in addition SNIA It is also proposed that Software definition storage (SDS) It should have the following characteristics ：

• High degree of automation – By simplifying management , Reduce the O & M cost of the storage infrastructure  
• Standard interface – Support API management 、 Publish and operate storage devices and services
• Virtualized data paths – Support multiple standard protocols , Allow applications to store... Through blocks , File storage or object storage interface writes data
• Extensibility – The storage architecture has the ability to scale seamlessly , The scaling process does not affect availability and does not result in performance degradation
• transparency – Storage shall provide users with available resources and expenses for managing and monitoring storage

What is distributed storage

There is actually no clear definition of distributed storage , There is not even a uniform name , In most cases it is called Distributed Data Store perhaps Distributed Storage System.

Wikipedia gives Distributed data store Is defined as ： Distributed storage is a kind of computer network , It usually stores information in multiple nodes in the form of data replication .

The definition given in Baidu Encyclopedia is ： Distributed storage system , It's about storing data on separate devices . Distributed network storage system adopts scalable system structure , Use multiple storage servers to share the storage load , Use location server to locate storage information , It not only improves the reliability of the system 、 Availability and access efficiency , It's also easy to scale .

Although the definitions of distributed storage are not exactly the same , But one thing is unified , That is, distributed storage places data in multiple nodes , Nodes provide storage services through network interconnection . This is obviously different from the way that traditional centralized storage places data in a centralized way .

Server SAN What is it?

Server SAN The concept , The first is the WIKIBON （ A famous storage consulting community abroad ） Put forward ,WIKIBON Yes Server SAN Is defined as ： A combination of computing and pooled storage resources , Multiple storage media are directly attached to multiple independent servers for use . These direct attached storage devices (DAS) Over high-speed networks （InfiniBand Or low latency Ethernet ） interconnection , And ensure the consistency of storage through software .Server SAN Mixed storage media can be used , Including mechanical hard disk and SSD Hard disk, etc. . At present, the mainstream cognition of the market will Server SAN As SAN The architecture of , Only storage no longer uses dedicated hardware and controller architecture . But in Wikibon Define the ideal situation , Enterprise applications run on Server SAN On , High availability , Even super fusion products are classified as Server SAN, But because this concept is not common in the market , So I won't focus on it .

What is hyperfusion

Refer to the definition of hyperfusion in Wikipedia ：

Hyper converged infrastructure （hyper-converged infrastructure） Is a software defined IT Infrastructure , It virtualizes common “ Hardware definition ” All elements of the system .HCI The minimum set that contains is ： Virtual computing （hypervisor）, Virtual storage （SDS） Virtual network .HCI It usually runs on standard commercial servers .

Hyper converged infrastructure （hyper-converged infrastructure） And Converged infrastructure （converged infrastructure）  The biggest difference is , stay HCI Inside , Both the storage underlying abstraction and the storage network are implemented at the software level （ Or by hypervisor Level realization ）, Not based on physical hardware . Because all software defined elements revolve around hypervisor Realization , Therefore, all instances on the hyper converged infrastructure can jointly share all managed resources .

Differences and connections

Software definition storage (SDS) The focus is on storage resource virtualization and software definition , First, in form , Software definition storage (SDS) Different from traditional “ Hardware definition ” Storage , It does not rely on proprietary hardware , The storage software can be run on a common server , It can avoid hardware binding and effectively reduce hardware procurement costs ; Have standards API Interfaces and automation tools , Effectively reduce the difficulty of operation and maintenance . Storage resource virtualization , Support multiple storage protocols , Consolidate enterprise storage resources , Improve storage resource utilization . But by definition , but Software definition storage (SDS) In terms of deployment form , Not necessarily distributed or centralized , in other words SDS Not necessarily distributed storage （ Although common Software definition storage (SDS) More distributed ）,SDS It is possible that the storage system is running on a single machine , Not storing data dispersedly through the network , This form of software defined storage (SDS) The extensibility of may have relatively large limitations .

Distributed storage , Its biggest feature is multi node deployment , Data is distributed over the network . Distributed storage is characterized by strong scalability , Load balancing through multiple nodes , Improve the reliability and availability of the storage system . And Software definition storage (SDS) contrary , Distributed storage is not necessarily software defined , It may be hardware bound , for example IBM XIV Storage , It is essentially a distributed storage , But it is actually delivered through dedicated hardware . Then there are still hardware bindings , High cost of ownership .

from Server SAN In terms of the definition of storage, the definition of storage is just like the definition of distributed storage , but Server SAN Emphasize the combination of computing and storage resources , That is to say, with software definition storage (SDS) and Where distributed storage is different , The later definition already has the rudiment of the concept of hyperfusion , But this concept has not become the mainstream .

The hyper converged infrastructure explicitly includes software definition storage from the definition (SDS), Have the ability of hardware decoupling , It can run on a general server . Hyper converged infrastructure and Server SAN The ideas advocated are similar , Convergence of computing and storage , Through a fully distributed architecture , Effectively improve system reliability and availability , And it has the characteristics of easy expansion .

SMTX OS Product architecture

Because many readers are confused about the composition of hyperfusion , The following SmartX Super fusion software SMTX OS As an example to illustrate the relationship between distributed storage and other modules .

Distributed block storage ,SMTX ZBS yes SMTX OS The core component of hyperfusion software . It adopts a fully distributed architecture and fully conforms to the concept of software definition .

SMTX ZBS Distributed storage architecture

besides , The hyper converged infrastructure is further extended , It emphasizes virtualized computing （hypervisor） At the core , Consolidation in a software defined way includes virtualized computing , Software defines storage and virtual network resources . From the author's point of view, there are more possibilities for the future of hyper converged infrastructure , It can promote the calculation , Storage , The Internet , Security , Disaster recovery and so on IT Service integration , Reduce IT The complexity of the infrastructure , Reshape ” Software Defined Data Center ”.

The technical core of hyper converged infrastructure

The hyper converged infrastructure can be said to integrate software definition storage (SDS), Distributed storage , Even Server SAN Many good features of , And it is not difficult to see that these concepts are inseparable from a keyword —— Storage . Obviously , The core technology of hyper converged infrastructure is storage , Store and host user data , The importance is self-evident , In addition, various advanced functions of super fusion （HA、 Online migration, etc ） Are heavily dependent on storage capacity , The expansion capacity of storage directly limits the expansion scale of super fusion products ; The manageability of storage also restricts the flexibility of the platform ; Considering the super fusion form , Computation and storage are the same software stack running in the server , The relationship between them is resource competition , Therefore, on the premise of ensuring the stability and performance of the storage software , The less you spend, the better .

Last , Summarize the storage requirements of hyper converged infrastructure , It can be used as a reference for investigating the storage capacity in hyperfusion ：

• Software definition – Unbind hardware , More functions can be expanded through upgrading , High automation capability
• Fully distributed architecture – Good scalability , Eliminate the risk of a single point of failure
• high reliability – Intelligent fault recovery , Rich data protection means
• High performance – Support multiple storage media , Take full advantage of the performance of new hardware
• Highly integrated – The architecture is simple and easy to manage