Cloud Computing Week Jevin Review (2.6-2.10)

File Transfer Protocol (FTP) will die

The File Transfer Protocol (FTP) is defined in RFC 959 and released in October 1985. File Transfer Protocol (FTP) is designed to be a cross-platform, simple, and easy to implement protocol.

File Transfer Protocol (FTP) has a long history of evolution, is one of the most important applications on the Internet, but today, has been declining. The author of this paper enumerates some shortcomings of File Transfer Protocol (FTP).

1. The data transmission mode is unreasonable

It is unreasonable to use ASCII mode to transmit data without considering the contents of the file itself.

2. The working method design is unreasonable

File Transfer Protocol (FTP) can work in active mode (PORT) or passive mode (PASV), which determines how data links are established.

3. Does not coordinate with the firewall work

Before the advent of File Transfer Protocol (FTP) in Network address translation (NAT) and firewalls, there was no malicious attack on the network at that time. Most end users today have IPV4 addresses that are not routable because of the use of firewalls and the shortage of IPv4 addresses.

What does this mean for FTP? This means that if the FTP client IP address is not routable or is behind a firewall, data transfer can only be done using passive transfer mode.

4. Password security policy is not perfect

In the early days of the Internet, File Transfer Protocol (FTP) did not provide password security. On the FTP client and server side, the data is transmitted in plaintext, and any person who has control over the routing on the communication path can sniff through your password and data.

5.FTP protocol Inefficient

Transfer a file, FTP need to reciprocating 10 times, and HTTP only need 2 times! If you transfer multiple files, FTP can omit the steps to send a username and password, and HTTP can transfer files in the same TCP connection using a fixed socket (socket).

AMD abandons X86 architecture development commitment to open processor instruction set

Recently, AMD abandoned its commitment to the development of the x86 framework, announcing a new design strategy to pave the way for the use of ARM technology in future AMD chips.

AMD says the company will allow the wisdom of third-party companies to be integrated into future processors as part of a tailored chip plan for consumers. If consumers need other architectures, including x86, AMD will combine its own design philosophy with Third-party companies to meet the needs of consumers. New AMD products will closely cater to the needs of consumers. Although the statement does not directly point to the use of ARM technology, the company is open to using other instruction set architectures, including Arm,mips and PowerPC.

MAPR Ceo:hadoop will enhance SQL rather than replace

Hadoop is almost synonymous with a relational data model, and SQL seems to be getting into the doghouse. John Schroeder, CEO of MapR, a well-known business Hadoop provider, believes that Hadoop will enhance SQL rather than replace it.

Last month, veteran IDC analyst Dan Vesset predicted that Hadoop would be a standard component of modern data centers. By 2015, the Hadoop market will mature, and the major players we know today may no longer exist. MAPR is a vendor of business Hadoop, the name of which is inspired by the MapReduce software architecture of Hadoop and is one of the companies on the Vesset acquisition target list.

So you can expect that MAPR CEO John Schroeder's prediction does not include his own company. But Schroeder at least sent a contentious argument: in the 1992 database market, the difference in 2012 was that the customer's new discovery capability led to the customer's tendency to turn from seller-led to restraining seller.

The next ten years: The Golden Age of Moresql

Moresql (previous newsql) is designed to combine the advantages of SQL and NoSQL. With technology integration, MORESQL can fully absorb the benefits of nosql flexibility and scalability on the basis of SQL (structured Query Language), ushering in the golden age of relational databases.

The rise of the NoSQL movement

The industry has long believed that the join command in relational database queries can adversely affect performance.

Although the ISAM storage engine in MySQL does not support transaction processing and fault tolerance, it is an unreliable storage engine, but it is favored because of the extremely fast access to the ISAM storage engine and the price advantage of MySQL. However, the cap theorem proves the doubts in everyone's mind that relational databases lack extensibility.

Google publishes a series of papers on MapReduce, distributed data storage, and scalability issues. Then Amazon joined the discussion. Soon after, everyone was involved in the uproar.

Tenzing Sql:google SQL Implementation on MapReduce framework

Tenzing is a query engine built on the MapReduce framework that Google uses for specific data analysis. Tenzing supports near-complete SQL implementations through scaling and consolidates a range of key features such as high performance, scalability, reliability, low latency, fragmented storage, structured data support, and more.

Google runs SQL on its own distributed data store and deals with tricky issues such as join in a special way. More than 1000 employees within Google Now use Tenzing to process more than 10,000 query requests per day over 1.5PB of compressed data.

Why Choose Moresql

Although NoSQL has greater flexibility and scalability, it has its own limitations. Not using Sql,nosql lacks a highly structured query that is tightly coupled with mathematics, and NoSQL does not support acid (atomicity, consistency, isolation, persistence).

Moresql (previous newsql) is designed to combine the advantages of SQL and NoSQL. With technology integration, MORESQL can fully absorb the benefits of nosql flexibility and scalability on the basis of SQL (structured Query Language), ushering in the golden age of relational databases.

The next ten years: The Golden Age of Moresql

I think the next 10 years is certainly the decade of the web, but it's also the 10 years we've been on the expansion issue to recognize the join. Over the next decade, we'll rejoin the SQL Renaissance (Moresql movement)!

Of course, Google will certainly publish new papers during this period, which may lead us in different directions. It's a risk, but it's worth looking forward to!

Subversion of the data center? Venture capital Nicira Tens of millions of investments

The enterprise's it construction is moving toward the more and simpler direction, the emerging cloud computing Technology service provider, Nicira, is dedicated to changing the way services and storage are virtualized. There is no doubt that this is a much-watched start-up. According to Nicira, he received a joint investment of $50 million from Andreessen Horowitz, Lightspeed Venture and New Enterprise Associates. Meanwhile, individuals including VMware co-founder Diane Greene and Benchmark Capital, Andy Rachleff, also invested in Nicira.

Martin Casado,nicira co-founder CTO

Nicira's NVP (receptacle virtualization Platform) is a software-based system that creates virtual distributed network infrastructure in the cloud data center and is completely detached from and independent of hardware facilities. Nicira says NVP is a flexible and intelligent way to turn hardware into software, simplifying and virtualizing the process.

The NVP platform creates a layer of software with IP panels on the hardware network that enables the creation of virtual networks that are indistinguishable from physical networks for security, QoS policy, and usability.

[Blog Highlights] LinkedIn senior Expert: My road to NoSQL

Recently LinkedIn distributed Data System team member Siddharth Anand on his blog titled "The State of NoSQL in 2012" article. It mainly expounds its own NoSQL application experience, at the same time expresses his view to the current enterprise NoSQL application, he thinks that the clarity and quality of code is very important, and must adhere to its own design principle.

Global Data Center Market pattern Change (Fig)

At present, the Global data center market development began to slow down, but the BRICS market still maintain rapid growth, data center construction work has just entered an active period. Meanwhile, developed countries such as the United States and Canada have been forced to reduce data center construction projects and consider how to improve data center efficiency. Into the 2012, the Global data center market pattern is quietly changing.

The timely revision of the United States strategy

According to Juniper NX Company data, the U.S. government plans to reduce the highly inflated IT infrastructure investment. Nearly 60% of business participants in the survey noted that their data centers used more than 20 different operating systems, and 16% of the business representatives said they used more than 100 operating systems. The same problem was encountered in the software sector: 48% of business representatives emphasized that they used more than 20 types of software, and 6% said that more than 100 software was in use. In this case, if the national institutions are not determined to replace the software, the transition to "cloud" and the reduction of the number of data centers will not only have a significant effect but also increase the burden on system management.

Europe attaches importance to energy saving

More than 8% of European companies plan to expand their data center infrastructure this year and next, according to a separate survey conducted by Digital Realty Trust, a UK data center service provider. The report says more companies are starting to measure the efficiency of energy use in data centers.

The survey, commissioned by Digital Realty Trust, was conducted in response to it decision makers in 205 influential European companies. 82% of the respondents said they would expand the capacity of the data center in the near future, and the survey also found that the UK became the most popular data center construction destination. 37% of respondents planned to build data centers in the UK, followed by France (30%), Germany (26%), Spain (21%) and the Netherlands (21%).

Russia to catch up

According to Datacenter Dynamics, the Russian data center market will surpass India, China and Brazil over a 2011-2012-year period from an investment point of view. In terms of growth in the data center, Russia is second only to Brazil in the BRIC countries.

The "cloud" is the protagonists

Cisco has pointed out that the fastest growing data center area is cloud computing. At present, the proportion of cloud service traffic in the data center total flow is about 11%. It is expected that by 2015, the annual total flow of data center will increase 4 times times, to 4.8 Chak bytes (zettabyte), the annual rate of traffic growth reached 33%. As a result, cloud computing plays a vital role in the future of information technology, as well as in video and content transmission.

Amazon cut cloud storage price per GB monthly fee of 0.125 dollars

Amazon's Network services department lowered its data storage price for its simple storage service (S3). The company announced in Monday that after a price cut, users could save about 12% of the cost per month by storing 50TB of data.

Amazon Cloud services charge S3 users monthly storage space charges. Now, Amazon has cut storage prices in the United States, when users store more than 1 TB of data, the monthly fee from 0.14 U.S. dollars per gigabyte down to 0.125 U.S. dollars, and then TB data monthly storage charges from 0.125 U.S. dollars per gigabyte to 0.110 U.S. dollars. If the data storage is increased further, reaching the 450TB and 500TB levels, the charges will be lower.

Large Data Technology Market research report: "Bigdata wave" forces enterprises to make choices

According to the IDC survey, global electronic device storage data will be 30 times times as fast as 2020, to 35ZB (equivalent to 1 billion 1TB of hard disk capacity). The arrival of large data waves has also brought a new challenge to the enterprise. For the prepared enterprise this is undoubtedly an information gold mine, can reasonably transform large data into valuable information to become the necessary skills of future enterprises.