Oracle Helps Linux Get Enhanced Data Integrity
Page 1 of 1
When you click "Save" for a file, you expect that the file will be saved. Unfortunately, due to a variety of reasons, that doesn't always happen.
If you're a Linux user, though, you might now have better success than other operating systems' users, thanks to new, block I/O data integrity infrastructure code from Oracle that's become part of the mainline kernel.
The code is in the currently shipping Linux 2.6.27 kernel, which debuted in early October. And as a result, its developers say it could dramatically improve protections against corruption during saving.
"When an application is reading or writing against storage, one of the things that could happen is silent data corruption -- 'silent' meaning that the application reports that data as being correctly written, but it's not," Monica Kumar, Oracle's senior director for Linux, virtualization and open source product marketing, told InternetNews.com.
"Silent data corruption can happen more frequently than you think, and it could cause millions of dollar of damage if it's in a mission-critical business," she said.
Kumar also said the enhancement also means that Linux now has an advantage that no other operating system currently offers -- a benefit that could serve the open source OS well as its distributions continue to compete with established proprietary systems.
"Because Linux is now the first operating system that has the T10 standard implemented, I think it does have a lead in that sense," she said. "I'm not saying other operating systems will not follow, but for now, Linux is leading in this space."
Building on a standard
The new block I/O data integrity infrastructure code is an implementation of an industry standard, the T10 Protection Information Model designed for end-to-end data integrity for enterprise storage systems.
The code also adds what are known as Data Integrity Extensions, a technology co-developed with storage vendor Emulex that enables integrity information to be transmitted from memory.
Kumar explained that the data integrity feature works as an additional checkpoint at the operating system level, verifying the data being written or read. As a result, the process is designed to make data corruption less likely.
"The validity of I/O operations is assured through the exchange of verification information during data transmission," she said. "The key is creating integrity metadata, so as the data travels the data path, the integrity metadata is validated."
According to Kumar, Oracle and Emulex were working on the data integrity feature since mid-2006. In July of this year, the code landed in the Linux kernel development tree, ultimately becoming included in the final 2.6.27 kernel release in October.
Currently, the technology works with the Linux-based Ext3 filesystem, though Kumar noted that work is ongoing to validate the data integrity technology against other filesystems as well.
At the same time, Oracle and Emulex are pushing an early-adopter program to get a small group of end users testing and deploying the technology, which could help demonstrate its real-world potential.
Since the data integrity technology is now part of the mainline Linux kernel, it is available to all Linux kernel users -- and it could end up benefiting others beyond just Oracle and Emulex. Among two of the likely beneficiaries are Linux distribution vendors Novell and Red Hat.
"It goes without saying that any technology that improves data integrity is valued by Red Hat Enterprise Linux customers," Nick Carr, marketing director at Red Hat, told InternetNews.com.
Michael Applebaum, senior product marketing manager at Novell, is also enthusiastic about the new data integrity technology.
"We see strong customer demand," Applebaum told InternetNews.com. "Our heavy database users ... for example, data warehousing vendors like Teradata, will benefit from enhanced data integrity."
"We think such advanced features further establish SUSE Linux Enterprise Server as the operating system of choice for mission-critical data center workloads such as major database, middleware and line of business applications, backed by our strategic partners SAP and Microsoft," he added.
For the present, Oracle is talking up its contribution as it and Emulex build an early adopter program to get the technology more widely deployed among storage vendors.
As a result, Kumar also suggested that storage vendors will soon be announcing gear that is specifically optimized for the technology, but provided few additional details.