Interview multiple backup CTO Chen Yuanqiang: Full Open Enterprise Data Cloud eternal life Road

Source: Internet
Author: User
Keywords Cloud computing public cloud Chen Yuan Strong multiple backup
Tags .mall access aliyun analysis application backup bandwidth basic

The advent of the 4G era, enterprise data faced with explosive growth, mobile TB; At the same time, human factors, software defects, uncontrollable natural disasters and other security problems occur frequently, how to make the enterprise data security and reliable, low-cost and efficient long-term preservation, has become an urgent concern for any enterprise. Fortunately, the cloud era accompanied by the 4G era, the core advantages of cloud computing: cost-effective, resource allocation, infrastructure flexibility, business smooth switching, bandwidth and storage unlimited expansion features.

Multi-backup cloud backup, cloud recovery, cloud archiving and other features can make full use of these high-quality cloud infrastructure to achieve enterprise data management services; Multi-backup is the first Internet company to focus on enterprise data cloud Backup, recovery, archiving, storage and migration, after 10 months of continuous iterations, Around the enterprise data security management services have been online, has been more than 40,000 of SME customer recognition, but also introduced internationally renowned capital investment. This time we contacted the multi-backup co-founder and CTO Chen Yuan, and let's look at how multiple backups are resolved to reliably store backups in the cloud. The following is an interview transcript:


CSDN: There are actually a lot of powerful people in the field of data storage backup, how are multiple backups considered for storage backup?

Chen Yuanqiang: Very good question! As we all know, there have been a number of major failures, including rich owners, banks and cloud giants. We all know that some of the critical systems ' disaster-recovery backup schemes are actually quite expensive, the program moves the billion, the annual maintenance fee is also high, the previous time in the online discussion of Ningxia Bank is because the cost is too high, and abandoned some maintenance plans, and finally led to a disaster. So the root of the problem is the cost. And the very low investment can solve the problem, which is one of the core advantages of multiple backup. In addition, the traditional program in the use of extremely complex, the general need for a strong professional ability to get started, and minimalist design is also a multiple backup changes in the traditional storage backup key, users only need 3 steps to complete the backup.

CSDN: Currently, including Amazon, Aliyun, EMC, Microsoft Cloud, Telecom, mobile and other giants have their own storage backup solutions, multiple backup and these cloud storage backup services, what are the characteristics?

Chen Yuanqiang: First, we are partnering with these cloud platforms or storage giants and we are a SaaS application of the architecture on these platforms, and we are both their customers and the core application of many of their platforms. Only a sufficient number of applications derived from these platforms, it is conducive to the Giants to build cloud eco-chain industry, so multiple backup is a core part of the cloud ecological chain.

From the beginning we are in the direction of the service on these mega-platforms, not their own development of the lowest level of storage, really through the integration of multiple backup computing, storage, backup and other services, so that the potential of these platforms to play out. Users can choose according to their own needs or we recommend the underlying cloud computing and storage platform, one of the key technology is Cloud 5, which is to block the data, and decentralized redundancy stored on different cloud platforms, to avoid Tan Yun failure caused by data loss.

EMC is a very powerful data management Service company, and experience in big data management is something we have to learn from, and EMC has been in the cloud area long ago, and has been a strong addition to the super mass storage backup scenario. Multiple backups are more about learning EMC's big data management practices and blending in with our innovative direction.

CSDN: Multi-backup positioning is very clear: the integration platform to form a strong aggregation capacity, into a user needs, according to the need to freely schedule the global cloud computing platform.

Chen Yuanqiang: Yes:-), indeed. For the start-up enterprises, do their own good things, with a wide range of platforms to form a strong cooperative relationship, is the easiest way to effect the road!

CSDN: How long does it take to consolidate these cloud platforms, and how much is the resources involved?

Chen Yuanqiang: Website from last October on line, currently lasted 10 months, in research and development on 13 people, our technical team are senior personnel, in the system, network, storage, security and other average have 7+ years of experience, mainly from Tencent, Shanda, Thunderbolt and other domestic front-line internet enterprises. Product daily Iteration on-line, the user discovered the problem or the flaw, basically is the day immediately changes, the weekly will have the big characteristic or the optimization on-line, accumulated down the small detail modification, hundreds of big optimization and the new function development. Products are now mature, whether it is large to TB enterprise core application data, or small to a few MB small web site data storage backup, with multiple backups are easy to complete the backup task.


CSDN: General storage Backup enterprises in the research and development of the resources invested in a large, more than 13 people spent 10 months of resources, specific ideas have some to share it?

Chen Yuanqiang: It is true that some of the existing traditional equipment or software resources are very large, moving hundreds of people. In fact, our idea is simple, the vast majority of products are either functional list or excessive design. Take the current CDP (continuous data protection) products, generally claimed to be real-time, 0 recovery window, but recently a view is that the real majority of enterprises in 15 minutes to recover the data can meet the requirements. Sometimes, in order to reduce the 1 minutes, you may put in the resources and time will be several times, so put the resources to the most necessary features, so that can actually solve 99% of the scenarios. This is also the Internet for product services, "Less is more" an application of the idea. Therefore, the idea of multiple backup is very simple, is focused on enterprise core data, such as various types of business (web, OA, mail system, file sharing system, etc.) server (LINUX,WINDOWS,AIX) file system backup storage, as well as common database (such as: Mysql,oracle, MSSQL, etc.) data backup storage, specific applications can support different data levels of the scene.

CSDN: Is it possible to share some specific ideas about multiple backup to achieve multiple storage backup of cloud storage platform data?

Chen Yuanqiang: This question is divided into several points to answer

1. Development of existing cloud computing platforms

Foreign Public Cloud basically formed AWS Reign, Google, Microsoft first Line cloud service coexistence situation, including RACKSPACE,HP,IBM,EMC and other traditional it vendors, mixed cloud platform coexist situation. Domestic Aliyun, telecommunications, Tencent Cloud, Ucloud, seven cows, Huayun and the old IDC manufacturers such as the western data, new nets, million nets, such as the introduction of orange cloud Platform, very much. Therefore, the number of foreign and domestic cloud platform is quite large, more domestic, cloud computing currently in the domestic time is only 3-4 years time, all aspects of continuous progress, performance, stability, service and so on.

2. What kind of cloud platform capability does the enterprise need?

with the information technology into every corner, the enterprise from the high profit center, into a small profits but quick turnover model, the vast majority of enterprises need is easy to solve business needs, easy to communicate with external low-cost programs. The advent of cloud computing platforms coincides with this trend. At the same time, also put forward a higher demand for the cloud platform: high cost-effective, stable and easy to expand, business flexibility, reliable data security. Enterprises can focus on their own business, integrate the use of high-quality IT resources for business services.

3. How multiple backups integrate these platforms and provide these capabilities

Basically, these cloud platforms have their own characteristics, the most direct case of multiple backup is the combination of their characteristics, the integration of its advantages, provide a simple direct data storage backup using the portal.

in foreign countries: AWS is the first public cloud platform, the scope of business coverage in addition to China other regions, platform maturity is very high. But in fact, if the domestic friends directly to use, will actually encounter a lot of habits problems, including billing, storage and the use of the host. Google is mainly individuals and gae, the enterprise-level IaaS platform has only recently begun to force, and Microsoft's platform relatively new, including individuals and businesses, stability is relatively not AWS stability, and some other restrictions are unique. Rackespace, the second-tier cloud computing enterprise, is relatively more difficult to base on the product experience. Foreign relatively good advantage is the basic configuration is better, especially the network bandwidth this block. In terms of domestic: product maturity, especially the reliability and stability of several mainstream cloud platform than foreign, the relative advantage of the place lies in the use of products more in line with the habits of the people, the biggest advantage is the service, encounter problems, you can find someone. Foreign products The biggest problem is communication difficulties, in addition to linode/digitalocean this kind of focus on cloud host services enterprises, response quickly, the other several giants of the product settings can not find the service entrance. Domestic IaaS enterprises are also gradually moving overseas, Hong Kong is the 1th station to connect the Southeast Asian market. But domestic relative abroad, in the bandwidth billing is the biggest difference, the cumulative flow of foreign billing methods, more suitable for multiple backup business characteristics. As far as the overall fault tolerance and equalization ability is concerned, these platforms do not realize the complete interconnection between the physical areas of the platform, and each service center is relatively fragmented. Foreign products have some trans-regional disaster-tolerant capability, and because of the good basic network, the impact of fragmentation is relatively not large.

So, in fact, a lot of cloud platform, but in product design, billing ideas, infrastructure, as well as the main direction of the strategy are unique. For companies, while abandoning traditional schemes, the business and important data that you want to put on the cloud can be as reassuring as the bank's money, not because of the cloud, but with the energy to take into account compatibility, usability, reliability, and security beyond the business.

For multiple backups, we are most concerned about the cloud platform three things: first, there is no high-performance host services, we need to back up the enterprise data processing, is an I/O operations-intensive business; the second is the reliability of storage, which is our lifeblood; the third is the distance between host and storage, user and host, Directly affect the speed experience of users using services.

So we focus on the consolidation of storage backup, through consolidation, providing a simple Easy-to-use interface, 6 9 (99.9999%) of data security reliability (such as our Cloud 5 technology), and providing a higher cost-performance ratio than a single cloud and traditional offsite data backup, Real data is never lost with little cost.

CSDN: Which countries are the areas covered by multiple backups?

Chen Yuanqiang: Multi-backup currently supports North America, China, south-east Asia, parts of western Europe, and other areas are currently covered by these cloud node areas, involving AWS, Microsoft Azure, Linode, Rackspace, Aliyun, Ucloud, Mobile cloud, telecom Cloud, Dropbox , Microsoft one Drive, Google Drive, Baidu Cloud and so on. Multi-backup current global area coverage distribution diagram.


CSDN: We have talked before, multiple backup both support the domestic, but also support the foreign cloud platform. How these platforms are currently seamlessly integrated into multiple backup platforms, especially given the fact that the bandwidth of international exports is actually very congested.

Chen Yuanqiang: For the integration of international and domestic platforms, at first we did not consider so much, more is the business model to make, so the early encounter a lot of problems, trans-International Line control command transmission loss, the domestic and foreign data around the transmission, scheduling also did not consider the distribution of the task area, Plus all kinds of cloud platform in fact the stability of such or such a problem, so the first challenge is very large, the task often card, or storage failure rate is high. The internet has a saying in architecture design, architecture is never designed, is constantly integrated operation optimization out; After nearly 10 months of continuous optimization iterative, product experience on a large step, in the user interface highlighting heuristic guidance, feature focus on the backup and restore the core experience of these two, The basic resources have done a lot of system optimization and regional routing balance, for the task of hierarchical partition scheduling, faster, more stable task execution.

CSDN: There's a security issue, and our previous interviews with other cloud services all mention the security of the cloud platform, and companies are very concerned about it. Is there any concrete measure to this multiple backup?

Chen Yuanqiang: Yes, security is the basis of multiple backup operations, but security is always 1 unfinished topic, multiple backup in security has been very important, reflected in the following several basic things:

The commitment to protect user data is clear from legal and internal regulations. System administration permissions are assigned by the person responsible, reducing the potential risk due to decentralization of authority. All service periodic security checks and risk assessments, and regular security policy adjustments. System Operation data Hot-standby and Lengbei, the system itself has the ability to recover. Sensitive data for all users is encrypted using AES algorithm, and for user backed up data, high strength encryption is done before the user's host is out, and user-related behavioral actions automatically perform security audits. For a task that uses cloud 5, it will be disorderly to break up in different cloud and storage facilities based on the mainstream cloud platform, multiple backup itself will enable the cloud platform itself security services, further enhance the system's risk discovery and prediction capabilities

Nevertheless, we hope that the whole industry will have more positive energy superposition, the real focus on the team's efforts to continuously promote the efficiency of the industry and respond to natural disasters, software defects or artificial misoperation caused by the failure to respond to the direction.

CSDN: We talked about some of the data scale issues, multiple backups can be global, fluent support from a few MB, to several terabytes of data storage backup, this piece can be introduced?

Chen Yuanqiang: First of all, the highly efficient and intelligent backup network architecture model of multiple backups, which is based on the design model of Linux kernel, is a model of layered large asynchronous and small synchronous drive design, each layer provides the core capability agreement for the upper layer.


In order to achieve TB/MB data can be mixed in Backup network fast backup, specifically from a number of links to start:

Data Access Layer

The data access layer is designed to be simple, with the easiest and most universal access to data, so that data from all kinds of scenarios are easily accessible to the backup system, while maintaining the minimum amount of data transmission at a time. Therefore, in addition to conventional block, compression, and other conventional technology, but also combined with the Internet's small bandwidth features, support for multi-level cache acceleration technology, with the most rapid differential analysis technology, so that every backup to keep the smallest amount of data into the transmission layer.

1th Time: Blue block represents the function block that the main function of the data is minimized, the figure indicates from 10GB to 1-5GB, the final output data also is related to the characteristic of the data itself.


After the 2nd time: on the 1th basis, the local cache-related speedup and variance analysis began to work, the actual changes in the data may be very small, the figure indicated that although the new 1GB, but the actual change in data only 0.1GB.


The current specific deployment form, we support Plug-ins, hosting, as well as client-side proxy mode, plug-ins mainly for small and medium (1GB below) Web sites, hosting support a slightly larger 1-30GB, while the client mode supports 30GB above data backup.

We currently have complete support for Linux (32, 64 bit), Windows (32,64), and Aix. Database currently we support mainstream mysql,oracle,mssql,postgresql. In the specific application scenario, we have a set of corresponding scene recognition capabilities, at the same time can be combined with the API, the control panel embedded in the cloud platform, virtual host manufacturers, as well as online SaaS services.

Data transfer Layer

At present, backup network work strategy: The basic is to use transmission and control to separate the way to design the overall transmission network, similar to the current SDN thought: Through the target type recognition, server partition coding, resource status and network state dynamic analysis, support the global intelligent scheduling ability, for different scales, Different areas of the object, we will automatically dispatch to the most appropriate nodes and lines to ensure optimal configuration of network nodes and bandwidth resources, optimize the backup recovery experience.


Specifically divided into several points to consider:

Node: Directly using the optimized OS bottom stack, quickly start the transmission window, speed data transmission speed.

Network: On a specific network deployment, by adopting the high quality node, the core transmission channel is set up, the node priority partition is set according to the data characteristics, the large and small data and the VIP channel are isolated, and the nearest access and Link aggregation strategy is initiated to improve the bandwidth utilization among the nodes. For some scenarios that require a higher level, For example, large scale data, we will also launch a regional acceleration strategy to complete the rapid transfer of segmented data.

Data storage Tier

At present, multiple backups support enterprise-level and personal cloud disk storage access, and support full redundancy grouping and cloud 5-block cross Cloud redundancy mode. The data is stored in principle, and the Cloud 5 mode is enabled to enter the Zone Acceleration mode transfer.

Improve data storage efficiency. In addition, in the cloud storage here, we have with several platforms have in-depth cooperation, than the general access to open API capabilities will be more powerful.


Storage for multilevel index design, support across the region distribution, support cloudy storage, while storing objects into blocks and file object combinations, providing high-speed storage, download, search and other capabilities.

Optimization is never over, we are currently working with top speed manufacturers in the country and more cloud computing and cloud storage enterprises, in the near future, will fully start the global data transmission channel acceleration, is expected to the current backup network on the basis of the current increase of 3-5 times the transmission capacity, Response times for actions such as backup and recovery, migration, and so on, for more than terabytes of data are significantly reduced.

CSDN: In recent years, cloud-related platforms and services have grown very fast, it is said that more than 40,000 companies in the use of multiple backup, blessing multiple backup in this wave of homeopathy to establish a reputation, and constantly launch the cloud on the surprise of the data management function, for Oberyun Enterprises to provide more pro-people data security services, so that data eternal life!

Chen Yuanqiang: Thank you, this is also our goal, we must work hard!

(Zebian/Wei)

Free Subscription "CSDN cloud Computing (left) and csdn large data (right)" micro-letter public number, real-time grasp of first-hand cloud news, to understand the latest big data progress!

CSDN publishes related cloud computing information, such as virtualization, Docker, OpenStack, Cloudstack, and data centers, sharing Hadoop, Spark, Nosql/newsql, HBase, Impala, memory calculations, stream computing, Machine learning and intelligent algorithms and other related large data views, providing cloud computing and large data technology, platform, practice and industry information services.

Related Article

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.