Let the cloud easily analyze big data

Today, more and more popular cloud concept today, can we imagine their own data stored in the cloud, do not need to move data in the case, you can make the data for complex queries and analysis? It's also Joyent, a high-performance cloud computing infrastructure and big data analytics company that is trying to solve problems with its new Joyent Manta storage service.

The goal of Manta , the next-generation cloud computing object storage and data services platform , is to bring computational and analytic capabilities directly to customer data in the cloud.

Bryant Cantrill, Joyent's vice president of engineering, who spent three years concentrating on manufacturing Manta, said "the killer app for personal computers is a spreadsheet," because the spreadsheet combines data and calculations, allowing business people to put data Run on the desktop without the need for time-sharing programs to do this

Cloud computing storage combines two pillars of technology, data and networking, and Manta goes one step further.

Is Manta a cloud computing spreadsheet?

"I think Manta is a spreadsheet in cloud computing," Cantril said. "It opens the door to the analysis of unstructured data. What we are doing is integrating computing with data and networking, The integration of these three into a single facility, all in one single product, will form the Holy Trinity of computing, network and data. "

Cantrill said that by eliminating the need to manage infrastructure and move data, Manta enables businesses to process big data faster and easier, while keeping the data secure and priced in line with the Amazon public cloud.

"You really can not get this product to market if your product is not price competitive," he said. "We think there's more to be competing than price, cloud computing, but at the beginning You can not overestimate the price of the product.You pay the price of S3 for storage and you will pay for the EC2 or Joyent Infrastructure as a Service for any calculation used to speed up your data, One second, two seconds, or three seconds, if that's the case, you pay for just a second, two seconds, and three seconds. "

For example, you may want to periodically verify your backups. You do not usually do this because it will mean dragging all your backups back into calculations. However, with Manta, this validation test will be very straightforward and you will only have to start the calculations on the backup (they will only be a few subtle).

Manta lets you perform many computational tasks, including log analysis, search index generation, financial analysis, and other data-intensive tasks without the need to move data or set up compute clusters and processing software. Cantrill explained that the code was introduced in parallel to the physical servers within the secure container, while the data was automatically merged through the industry-standard MapReduce mode.

No need to copy data from storage to compute cluster

Konstantin Gredeskoul, CTO of Wanelo, an online shopping community, says it can take hours to copy data from storage to compute clusters.

Gredeskoul said: "The Joyent Manta storage service allows businesses to move closer to instant queries and analysis without the need to move data, seamlessly and cost effectively, and we're now able to do hundreds of gigabytes of data in minutes Queue analysis and generate a retention report, which is quite revolutionary when you compare this with traditional methods such as data warehousing. "

"With 50% of the world's smartphone traffic coming through Ericsson, we are constantly evaluating new technologies to increase the network's ability to manage the growing data while ensuring security, confidentiality and cost-effectiveness," said Ericsson Corporate Strategy and Marketing. Vish Nandlall, chief executive officer and CEO, said, "Joyent's new compute-storage innovation is a fundamental paradigm shift that changes the economics and utility of object storage and high-performance big data analytics."

Joyent Manta Storage Services provides the following features:

Multi-datacenter object storage with fine-grained replication control

No object size limit

Very consistent writing and highly available reading

Each object replication strategy

File system-style namespaces, including directory queries

Manta's functionality comes from operating system-level virtualization

Cantrill pointed out that Joyent to achieve these functions mainly through the focus on OS-level virtualization, rather than focus on hardware virtualization. Operating system virtualization allows Joyent to initiate calculations on the internet-facing object store where the object is stored.

"If you need to support traditional operating systems, the virtual hardware is very powerful," he said. "But if you want to give your customers the highest level of performance you can not virtualize your hardware, you should virtualize the entire operating system. The program runs directly on the hardware and is in a secure and sealed manner. "

"You can effectively bring your analytics into the cloud," he added. "You can run whatever program you want on your data: Java, Python, Perl, or just Unix Shell. If you have Your own analysis tools or other tools, or if you have any analysis software, you can push them into the cloud and analyze the target data. "

Joyent has now partnered with Panzura, a data storage management company, to help customers securely migrate their data from existing NAS, backup and archive storage to Manta.