Nirvanix Shutdown: Small File Users May Fare Worst
Nirvanix customers face a looming deadline to move data out of the cloud service. According to Nasuni's cloud data speed testing, moving small and medium files will be painful.
Nirvanix has told customers they must stop replicating data to its cloud service and download any data they wish to save by Oct. 15, although it still hasn't made a public statement or changed any of the marketing or other information on its website.
The narrow window for downloading raises the question of how large Nirvanix's download pipes are and whether customers will be able to get production data out in time.
All Nirvanix customers received the following notice Tuesday, according to TechCrunch's IT Knowledge Exchange, an online forum of IT professionals: "We are notifying you as soon as possible after making this decision so that you can make alternative plans for storage service. Nirvanix will have resources available to continue to provide service between now and Oct. 15 for you to download your data free of charge."
In some cases, Nirvanix was being used as a storehouse for second or third copies of data, as an archival site rather than primary data storage. Those customers may just stop replicating to Nirvanix and replicate to a new site. They should also seek assurance that those backup copies will be appropriately destroyed as Nirvanix withdraws from the cloud storage business.
Customers who have unique or primary data sets on Nirvanix have a bigger problem.
Cloud services post what they charge for a given amount of bandwidth as customers download or move data between availability zones on Amazon Web Services, but they rarely say how much total bandwidth can be made available at any given time. The more customers who try to download data between now and Oct. 15, the more Nirvanix's existing bandwidth will be divided up, resulting in what may be slower data transfers.
Nasuni, an online service focused on guaranteeing data transfer speeds, attempted in 2011 to figure out how well various cloud services would perform, given the task of transferring small, medium and large data sets. Nirvanix was one of six tested at the time, although it was dropped from a subsequent annual report on data transfer capabilities because Nasuni became wary of placing its customers' data on Nirvanix after the first test.
Those test results may shed light on Nirvanix's present data-write capabilities. They also shed light on the data transfer capabilities of other cloud services that Nirvanix customers may be currently considering, such as HP, Amazon Web Services, Microsoft Azure, Rackspace, AT&T or Peer1. The figures indicate that Nirvanix stood with the best of service providers when it came to reading large data sets coming in. It ranked in the lower tier, fourth out of six, when it came to its speed of writing data, a limiting function when it comes to sending data out.
Nasuni compared the transfer speed of small (1 KB), medium (128 KB) and large (1 MB) files.
"Two cloud service providers emerged as top performers, Amazon S3 and Microsoft Azure, with S3 being the standout across all evaluation areas (performance, availability, scalabilty)," concluded the 2011 report.
Nirvanix made the cut in 2011, listed among the service providers that met Nasuni's minimal standards. Ten out of 16 services tested didn't, according to the Nasuni report made public on the tests.
InformationWeek obtained copies of Nasuni's 2011 and 2013 reports to see what they could tell storage users about Nirvanix data transfer performance and that of other cloud suppliers in 2011, and how their performance may have changed two years later.
Nirvanix outstripped overall standout AWS S3 in 2011 when it came to writing large, 1 MB files. Nirvanix was able to write close to 2.4 megabits per second, compared to Amazon's 2.1 Mbps, according to bar charts in the report. Only Microsoft was faster in the category than Nirvanix.
Likewise, Nirvanix excelled at reading large files, leading the pack at 13.3 Mbps, compared to Amazon's 11.1 Mbps and Microsoft's 13.1 Mbps.
When it came to writing medium-sized, 128 KB files, Nirvanix dropped back to fourth at 1.4 Mbps, compared to Amazon's 1.9 Mbps and Microsoft's 2 Mbps.
In writing small, 1 KB files, Nirvanix fell further behind proportionally, writing roughly 45 small files per second compared to Amazon's 129, according to InformationWeek estimates derived from bar charts printed in the report. Microsoft led the pack at 152 files; number three, AT&T, wrote 98 files; number five, Rackspace, 28 files; and number six, Peer1, handled 23 files.
The Nirvanix change in position illustrates that it was geared to write large enterprise files most efficiently compared to the other services, but ranked fourth on the writes of the two smaller file sizes.
The read function is also necessary in data export, and again the gap between Nirvanix and market leaders was greatest in the movement of small files. Amazon read 388 small files per second; AT&T, 270; Rackspace, 200; Microsoft, 153; Nirvanix, 127; Peer1, 47.
When it came to reading large files, the rankings were: Nirvanix, 13.5 Mbps; Microsoft, 13.4 Mbps; Amazon, 11.1 Mbps; AT&T, 9.9 Mbps; Rackspace, 6.2 Mbps; and Peer1, 2.6 Mbps.
Reading at a high rate of speed will aid taking large files into your service, although writes may back up in the process.
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.
Join us for a roundup of the top stories on InformationWeek.com for the week of December 14, 2014. Be here for the show and for the incredible Friday Afternoon Conversation that runs beside the program.