By Steve Lessem, July 11, 2010
This blog was reposted from the “Cloud Storage Strategy” website.
There is no doubt that every enterprise has devoted some time and energy to evaluating how cloud technologies can best be put to work in their ongoing pursuit of cost reduction and to a lesser extent for potential improved service levels particularly around rapid provisioning of compute and storage resources. Mezeo has recently begun to work with various enterprises, and I want to share some of the opportunities that appear to align strongly with these two goals.
In terms of cost, most enterprises are experiencing continued and significant growth in unstructured data. As they look at the cost of this growth, it is more than just physical storage, data center occupancy, bandwidth utilization and power and the accompanying management demands; it is also the backup and disaster recovery requirements and the ability to quickly satisfy users who need more storage in order to execute whatever tasks and jobs they have. Against this backdrop, the drumbeat of Amazon S3 and other public storage clouds advertises storage at costs that are generally below the internal “advertised” cost of the typical Fortune 500 company. What gives?
First, cents/GB/month is only the tip of the iceberg, and bandwidth along with access charges gives a more realistic cost appraisal. Next, real and legitimate concerns about data security exist (will someone gain unauthorized access, by accident or via an attack, to company data stored in a multi-tenant public storage cloud?). Also, data integrity concerns are well founded (will the bits I store be returned, and will they be backed up and appropriate DR measures taken?). Finally, can I absolutely trust the service provider to execute to the extent deemed necessary, and if they do, can they really save me any real money versus the assumed risk profile? Private cloud computing is an appropriate strategy for addressing these issues.
Not all unstructured data is a candidate for the latency of cloud storage as delivered from an IT service provider via the Internet. So, while some tiers of data may be appropriate for a cloud storage service, it is a subset of the enterprise unstructured data requirement and not a lower cost panacea. Hopefully, CIOs can easily make this case with their peers in senior management, although it may sometimes seem like they are making an excuse for keeping control and not exploiting new technologies.
Question one surrounds the cost proposition, and our analysis suggests that, even at sub petabyte initial cloud sizes, the enterprise can deliver economics for in-house cloud storage that compare very favorably. In fact, it may even be lower than what is available from a service provider. The Mezeo team comes from both a hosting and a cloud storage background, and this just reinforces our view that the cost proposition for private cloud storage has favorable economics. However, if you are being forced to allocate capital for data center build outs, or you are otherwise CAPEX constrained, the hosted public cloud economics can be quite appealing. Since businesses require positive margins, this further drives up the cost of cloud storage as hosted at a public service provider.
The case for improved user satisfaction is similar, regardless of public versus private, because the cloud gives users the capabilities they want. First, with rapid provisioning of pay-as-you-go low cost cloud storage, the end user gets what they need when they need it via a frictionless interface. Second, several benefits drive end user demand for cloud storage; including: avoidance of workstation storage upgrades, one solution for file sharing and collaboration, new capabilities and applications that exploit file search, tagging and publishing to a public URL, and the ability to access your storage anytime, anywhere and on any device. Third, the solution is also ideal for implementing a workstation backup solution with sync. It is not hard to see why end users would find all of these capabilities appealing.
Cloud storage clients, gateways and edge devices are also beginning to appear, and can solve many different issues. For example, a client gives the end user access to multiple cloud storage accounts at multiple providers. Why not replace that tape backup operation at a remote location with an iSCSI interface directly to a storage cloud, for a scheduled backup without local user intervention (get rid of the tape backup of your local file server, forever)? Speaking of file servers, multiple solutions for replacing or even displacing file servers are coming to the market. The savings from removing an entire layer of infrastructure are quite compelling.
New applications, including use of social media, may require file publishing. Cloud storage allows you to store training videos, and make them easily available at every end user in the company. Tagging and search offers new application capabilities, and new opportunities to support existing compliance requirements. Secure file sharing, versus file publishing, may be a significant requirement as you work with customers and business partners. Partner, customer and employee portals can reach new levels of capability with API accessible cloud storage, as the availability and the management of information is delivered via the cloud.
Our observation is that the early adopters have begun the move to cloud storage. Why? Simply, enterprise private cloud storage allows you to gain many of the benefits and set aside the security and data integrity concerns of public cloud storage. At the same time, data tiering and private and public could solutions will drive “hybrid” cloud approaches that will allow the enterprise to exploit the best of both worlds.
Trackbacks /
Pingbacks