- 1 VMware Takes the Wraps Off vRealize Automation and vRealize Business
- 2 Google Kubernetes Container Orchestration System Gaining Steam
- 3 VMware Covering Its Container Bases for Both the Short and Long Term
- 4 Containers Not Dealing a Death Blow to Public Clouds Anytime Soon
- 5 When Virtualized Infrastructure Comes at an Added Cost
- 1 Vapor IO Brings OpenDCRE to General Availability
- 2 VMware Takes the Wraps Off vRealize Automation and vRealize Business
- 3 Microsoft Previews Hyper-V Containers for Windows Server 2016
- 4 Mirantis Led FUEL Project Gets Installed Under OpenStack Big Tent
- 5 Red Hat Enterprise Linux 7.2 Adds Security, DR Features
Who Offers the Best Public Cloud Storage System?
One of the key uses of server virtualization technology is to enable public cloud computing, and one of the key uses of public clouds is to provide cloud storage. But deciding who offers the best public cloud storage system is a tricky task, as we shall shortly see.
One company that attempts to evaluate public cloud storage providers is Nasuni, a MA-based storage vendor. The company recently published a report entitled "The State of Cloud Storage in 2013" that looks at the public cloud storage offerings from five big companies involved in cloud computing using a variety of different hypervisors: Amazon, Google, HP, Microsoft and Rackspace.
Last year Nasuni found that the top offerings were Amazon S3 and Microsoft Azure Blob Storage, which may not be surprising: Amazon S3 is a mature and well-tested product, and Microsoft's Azure, while new, is developing fast.
The big surprise is just how successful Azure's development has been. That's because Nasuni's report found that Azure currently beats Amazon S3 in three primary tests: performance, scalability and stability.
"Even though Azure has not been in the market as long as S3 and does not have nearly the same volume of storage under management," the report states, "it is clear that Microsoft's investments in Azure are starting to pay dividends and that the technology they are providing to the market is second to none."
Two of the big changes that Microsoft has carried out are using SSDs instead of hard drives for journaling, and, probably more significantly, implementing a 10Gbps storage node network to replace the older 1Gbps one.
Improvements in reliability and throughput were also achieved by replacing a hardware load balancer with a software one. There were other changes as well, the details of which you can find on the Windows Azure blog.
The good news for anyone interested in public cloud storage is that Nasuni's tests also found that all the other providers had made significant advancements in their services over last year, including improved performance and fewer errors.
"It is clear that the minimum bar is moving upward, which is excellent news for the cloud storage market as a whole. As more CSPs mature into enterprise-class cloud storage providers, organizations and vendors will be able to leverage competitive advancements in price and technology to improve their overall storage infrastructure," the report concludes.
But the folks at Nasuni are a smart bunch, and even though Azure currently beats all others in the company's tests, it sticks to Amazon for its primary storage supplier, and Azure only as a secondary supplier in some cases. Why? Nasuni's Connor Fee explains over at GigaOm: "One major thing we evaluate is maturity and experience in the market, and Amazon still clearly has the most experience and is the most mature player in this space."
What happened next shows how prescient (or perhaps just lucky?) the folks at Nasuni really are. Azure suffered a global outage for around 12 hours. The cause? A simple schoolboy error: the Microsoft people forgot to renew an SSL security certificate. Oops!
And it's not the first time Microsoft has had certificate problems with Azure — something similar happened almost exactly a year ago.
What this all goes to show is that while the infrastructure that's put in place to support a public cloud computing offering is important, it can still all be let down by administrative incompetence or simple human error. And it's more likely to happen in environments that are less mature, and where administrators are less experienced in running it.
But before concluding that this says anything about public clouds in general being too risky for enterprises to depend on, let's not forget that human error can happen anywhere: enterprise IT people have forgotten to renew certificates before, and they will no doubt do so again. This time it happened to a public cloud storage provider, but that doesn't mean you shouldn't trust the cloud.
Paul Rubens is a technology journalist and contributor to ServerWatch, EnterpriseNetworkingPlanet and EnterpriseMobileToday. He has also covered technology for international newspapers and magazines including The Economist and The Financial Times since 1991.
Read more on "Cloud Computing" »