7 Comments

Summary:

Red Hat says its new POSIX-compliant virtual storage appliance will make it easier for IT shops to move legacy Unix applications to Amazon’s public cloud. The scale-out NAS appliance, based on Gluster technology, also replaces Centos with Red Hat Enterprise Linux.

5010167673_e796fa6c38_z

Red Hat’s new virtual storage appliance takes aim at a knotty problem for IT shops: Taking old-but-still-running legacy applications to Amazon’s public cloud.

When companies put workloads into Amazon Web Services (AWS), it’s usually all about new applications — older applications must be architected for the cloud and many companies won’t bother.

The Red Hat Virtual Storage Appliance for Amazon Web Services – based on Gluster technology — adds POSIX compliance. That means that it can take on legacy Unix applications. (POSIX is a standard that governs how applications and operating systems interact, assuring that an application written for one POSIX OS will run on other POSIX OSes. Most Unix applications are POSIX-compliant.)

“The cloud has not been POSIX compliant — AWS, AT&T, Microsoft are all great cloud companies but these older applications must be rewritten to use these clouds,” said Tom Trainer, storage product marketing manager at Red Hat.

The NAS appliance lets companies aggregate their Amazon Elastic Block Storage (EBS) and Elastic Compute Cloud (EC2) instances into a virtualized storage pool that can handle even these legacy Unix applications, Red Hat said. For this Red Hat-branded release, it also subbed in Red Hat Enterprise Linux for Centos as the core operating system. (Red Hat bought Gluster in October.)

Not all Unix is the same

But there is legacy and then there is legacy.  Red Hat is talking here only of Intel-based Unix applications, not Unix applications that run on RISC systems–which tend to be older. While newer web-centric companies may not see a need for this capability, there are still tons of these Unix applications running workloads, said Bernd Harzog, principal analyst with The Virtualization Practice. “The installed base of Unix servers across Solaris, HP-UX, and AIX on non-Intel platforms is massive. None of that stuff goes to the cloud easily,” he said.

Moving existing older applications to AWS will come down to a cost-benefit analysis, agreed Forrester Research analyst Vanessa Alvarez. “If Red Hat can make it easier, that’s great, but it depends on whether it makes economic sense to do so. Is it going to be resource intensive to move these apps or to re-architect them to fit the cloud? Or are organizations looking to move away from these apps [altogether]?” she asked.

Other companies are attacking the problem as well. This week startup Pneuron announced its take on the task of extending the viable life of legacy applications to private or public clouds.

Given Red Hat’s leadership in enterprise Linux, it makes sense for it to help customers more gracefully move their applications and extend their data center to the public cloud. Whether actually doing so makes sense will vary by customer.

Feature photo courtesy of Flickr user vaxomatic

You’re subscribed! If you like, you can update your settings

  1. So how exactly can I/O be guaranteed (latency/throughput of Disk)? Image an app having entirely random I/O responses based on the AWS Infrastructure layer, I have a feeling quite a few legacy apps would not benefit from this. Let alone attempting to troubleshoot performance problems.

    1. hi @mcrory i’ve asked Red Hat to respond to this — i’ll ck back to see the answer.

      1. I found a bit more info and see that if Clustering is done the I/O can be improved, however Network latency/throughput (there are no guarantees on AWS) along with the troubleshooting errors/etc. still seem to persist, so I look forward to hearing a response.

        Thanks

  2. I can’t see how this would work with a ‘legacy’ application that was built with scale assumptions in the code. The hardest part about migrating legacy now days is pulling all the scale out of the code/application.

    1. hi @mark, i’ve asked Red Hat to respond– i want to see the answer too!

  3. Hi folks, thanks for reading Barb’s piece and thanks for the feedback. Certainly, Red Hat Virtual Storage Appliance is not a one-size-fits all solution, however it will fit a great many. This solution has been deployed in production environments for approximately a year and customers are having success with it (see the CIC case study at http://www.redhat.com and more VSA case studies on the way). Barb’s included quote by Vanessa Alvarez is fair and accurate and each organization must make application choices that are best for their business. If they choose to move their apps to NAS in the AWS public cloud, then it is now a fully available and deployable solution. Tom Trainer, Red Hat Storage Marketing.

    1. I’m sorry, how does this answer any of the above questions? Please remove the marketing speak and address the questions asked.

      Thanks in advance.

Comments have been disabled for this post