Technology cloud report: East to West computing is not only about "computing", but also needs "new storage"

Original technology cloud report .

As a great change in the digital age ,“ Count from east to west ” Has become a “ Known to every household ” Issues .

at present , China's data volume accounts for 18%, The imbalance between supply and demand of computing power needs to be solved . Because of this , The state passes “ Count from east to west ” engineering , Layout construction “ New computing power network system ”, Promote the development of intensive computing . Many people will “ Count from east to west ” On a par “ south-to-north water diversion ”、“ West to east gas transmission ” And other vast projects related to the quality of people's livelihood .

However , stay “ Count from east to west ” Under the grand scene of , There are still many practical challenges .“ Count from east to west ” Project deployment , It will promote the east to count and the west to save （ Hot spot area data calculation 、 Western backup storage ）、 Count the East and train the West （ Eastern algorithm warehouse 、 Western offline big data analysis ）; For businesses with low real-time requirements, such as document cloud 、 Palm insurance 、 Development test and other applications are deployed in the Western node , Promote data center intensification 、 Cluster development .

From a storage system perspective , These segmentation scenarios bring new challenges , for example ： Local / How to access hot data in near field with high performance ？ How to flexibly migrate massive data ？ How the details of storage system layer operation are not affected , Present convenient business applications to customers ？ How to guarantee the security of data storage ？ How to meet “ Computer network symbiosis 、 Smart choreography ” Based on the architecture requirements of , Effectively reduce storage costs and energy consumption ？

Under the challenge of computing power network Mobile cloud “ New storage ” In time

stay “ Count from east to west ” scenario , Any cloud storage provider must face the challenge of storage system in the computing network , That is to solve the problem of high performance of storage system 、 High expansion 、 Easy to use problems .

Mobile cloud as China Mobile implements the National “ Count from east to west ” An important part of policy , Born at the right time, there is a way “ New storage ” technology roadmap .

Based on the operator cloud “ The Internet + Security ” Characteristics and advantages of , Mobile cloud is based on cloud network convergence , Build cloud edge computing power collaboration 、 Network convergence 、 Intelligent social computing service platform , Include “ Based on self-developed architecture , Create a three-dimensional ubiquitous computing power distribution , Unified heterogeneous packaging , Pull cross domain collaboration , Build more efficient 、 More flexible 、 More trusted computing infrastructure ”, Build a new storage capacity under the computing power network .

Based on this , Mobile cloud storage systems can “ Carry on 、 I can save 、 Good use ”： Meet the storage performance requirements of enterprise key business cloud , Provide high bandwidth 、 high IOPS、 Low delay guarantee ; Meet the fast-growing data storage demand , Provide easy to expand 、 High throughput 、 Efficient indexing of massive files ; Mask complex storage system level operational details , Provide easy deployment 、 Easy to operate 、 Simple and easy-to-use product form .

More Than This , Mobile cloud is also efficient and lossless 、 Infinite expansion and security of computing power , A cloud dedicated to computing power 、 edge 、 End global coverage , Rich public 、 private 、 Specialized multi category service patterns , Accelerate cross domain resource connectivity 、 Cross domain application collaboration , To be specific ：

（ One ） Efficient lossless and infinitely scalable

be based on DPU Uninstall to improve performance , Pull through bare metal 、 The virtual machine 、 Container, etc , Support smooth migration of applications , Realize the software and hardware collaboration of storage products , Calculate force unloading .
Through unified local storage 、 Distributed storage access framework , Combined with elastic lossless network , Build large-scale high-performance computing clusters , Achieve high-performance storage and computing collaboration .
Optimize load aware scheduling 、 Dynamic power management in the system 、 Ultra high density storage 、 Memory scheduling, etc , Realize green energy saving and efficiency improvement .
Build an efficient and agile cloud native infrastructure , Create a fast expansion capacity 、 Ability to deliver efficiently .
（ Two ） Computational security

Based on Trusted Computing and confidential computing , Realize the full stack security and trust from the underlying security chip to the key applications of the virtual machine , Improve the security protection capability of the computing power node . At the same time, it is based on data full link integrity verification , Ensure data accuracy and security .

Mobile cloud self-developed storage engine The whole system of storage products enables multiple business scenarios

“ Count from east to west ” Engineering highlights the value of data as a factor of production , And classify them for governance —— For service background processing 、 Storage disaster recovery 、 Offline analysis 、 Storage and backup do not require high network delay “ Cold data ”, Give priority to Guizhou 、 Inner Mongolia 、 gansu 、 Ningxia node transfer ; meanwhile , High demand for network delay “ Thermal data ”, Choose to decorate in the eastern city .

In order to fully adapt “ Count from east to west ” Multiple application requirements , Mobile cloud is based on self research “ Dayun Tianquan unified storage engine ” The superelevation of IO Performance and cross domain erasure capability , Promote distributed block storage 、 Object storage 、 File storage and other storage products are upgraded , Deep coverage of a variety of application scenarios .

Block storage products ： Based on Tianquan storage engine , Create millions IOPS、 Hundreds of microseconds delay 、4GB/s A new generation of cloud disk products with throughput , It can effectively support large databases 、 Real time log analysis, etc IO Intensive and AI Training 、 High throughput business scenarios such as gene sequencing , Meet the demands of on-site and near-field calculation .
Object storage products ： Through storage multi protocol collaboration (Storage Mesh), Realization “ Save and calculate in one , Count as you go ”, break “ Storage wall ”, connected “ data silos ”, make “ Data can be taken at will ” Storage service system .
File storage products ： Based on the cross domain erasure capability of object-based storage , Self developed mobile cloud file storage （CFN edition ）, be based on EOS Cross domain correction （ data ） And cross domain KV database （ Metadata ） Effectively support cross regional data sharing scenarios such as real-time streaming media file editing , Realize multi-point access 、 Simple configuration 、 Follow your heart .

Overall speaking , In terms of product system empowerment , Mobile cloud can realize data sharing across domains 、 Visit nearby , Support intelligent layering of hot and cold data 、 Flow on demand , Multi level disaster tolerance capability , Meet the sub millisecond access delay , Provide customers with high performance 、 High expansion 、 Highly concurrent data storage services .

As one of the underlying infrastructure capabilities , The deployment of mobile cloud cloud storage products has covered all resource pools of the mobile cloud . among ,“ Heaven power ” The deployment scale of storage series products exceeds 3 10000 servers , Capacity scale exceeds 2500 PB, service 80 More than 10000 customers . In addition, Dayun Tianquan unified storage engine technology is independently controllable , Adapt to localized servers , Realization “ One cloud, many cores ”, Truly safe and efficient .

Respond to “ Double carbon ” policy Mobile cloud storage implementation “ Energy conservation and emissions reduction ”

In the process of computing service , Massive data processing “ Energy consumption and carbon emissions ” The problem has also attracted much attention . In this , Mobile cloud response countries “ Double carbon ” policy , Relevant countermeasures have been considered when building the storage capacity of the new generation of computing network .

for example , Through the above mentioned “ Storage multi protocol collaboration ” Realization “ Save and calculate in one , Count as you go ”, structure “ Data can be taken at will ” Green energy-saving Data Center , And the intelligent data layering characteristics of storing multiple specifications and between multiple specifications according to objects , Let cold data flow gradually to lower storage costs 、 In storage types with lower energy consumption , Achieve energy conservation and emission reduction .

Specific to product performance , The fourth generation of mobile cloud object storage implements cross domain erasure correction , Make the data redundancy mechanism available from 3 Replica 300% Redundancy is reduced to 6+6 In the form of erasure codes 200% redundancy , Lower investment costs 33%.

future , With the blessing of the advantages of computer network integration , Mobile cloud will continue to innovate storage technology 、 Optimize the whole line of storage products , Give consideration to safety, efficiency, green and low carbon , by “ Promote balanced development between the East and the West 、 Improve the anti risk capability of the industrial chain 、 Decarbonization growth of enabling industries ” Contribute .

【 About the technology cloud report 】

Focus on original enterprise level content expert —— Technology cloud report . Founded on 2015 year , It's cutting edge enterprise class IT field Top10 The media . Authorized by the Ministry of industry and information technology , Credible cloud 、 One of the official media designated by the global cloud computing conference . In depth original report Cloud Computing 、 big data 、 Artificial intelligence 、 Blockchain and other fields .