Shifting to Software-Defined
Showing results for 
Search instead for 
Did you mean: 

Private Clouds – Cost Considerations of Proprietary vs OpenStack® Distributions


In the sixth part of this Hybrid cloud economics blog series, I discussed when you may want to create an in-depth TCO comparison between public cloud and private cloud alternatives to make better informed decisions. In this seventh part of the series, I will take an even closer look at the economic considerations proprietary versus open source distributions play when evaluating a self-managed private cloud solution.

Proprietary vs OpenStack private cloud

When deciding between public and private cloud for a specific requirement, there are times when a private cloud will be the best deployment option. Security needs, regulatory compliance, data sovereignty, and application architectural requirements are just a few examples of business considerations that can drive an implementation decision to a private cloud. When it comes time to deploying a private cloud solution, there are two general options: proprietary and open source. In the case of open source today, the primary option is a solution based on OpenStack.

Proprietary cloud solutions are built on closed source software that cannot be modified by the user. With a proprietary solution, the vendor takes on responsibility for all software development, maintenance and support. This can be done either through an annual license covering all of the above, or through a perpetual license for the software (object code), and annual payments for maintenance and support.

In earlier blogs, I covered many of the different factors to consider when comparing the cost of private cloud. Now, I also want to point out the impact vendor lock-in can have when evaluating a proprietary solution. Since there is only one commercial vendor for the software and the software will have vendor “hooks”–such as APIs–specific to the platform, you are likely to find it difficult to switch to a different solution in the future even if that other solution can be purchased at a lower cost. Applications may need to be re-architected and recoded. Data may also need to be reformatted and even physically moved from the original platform to the new platform. Costs of performing these actions could be prohibitive enough to move to a different solution.

By contrast, the OpenStack open source option offers greater flexibility for switching vendors later, should you want to do so. You may also find that applications can be more easily moved between clouds from different vendors, enabling you to take advantage of capabilities, such as cloud bursting when application demand exceeds capacity on any particular private cloud. Open source development draws from an industry-wide talent pool, which has the potential for offering more features at a faster pace than proprietary solutions. This has very much proved to be the case with open source operating systems (Linux) and databases. There is no reason that it will not happen with cloud technology, as well soon.

With an OpenStack solution, you have two general options. The first is to take a do-it-yourself (DIY) approach. This involves getting the OpenStack software from a community distribution, and then hiring staff to install, operate, and maintain the software by pulling bug fixes, patches, and updates from the open source repository. While OpenStack provides a highly functional cloud solution, at least at present, the general distribution may still lack specific capabilities that you require (For example, disaster recovery and high availability are not currently part of the OpenStack distribution). This means that you will either need to fill in these gap capabilities yourself and then integrate them in, or you will need to find a third-party vendor that offers what you are looking for and work with them to do the integration. In either case, this is likely to involve significant time, cost, and possibly development risk. However, the big benefit of the DIY OpenStack approach is that you won’t have to pay any license, support or maintenance fees. The software is free, as is support, which is provided by the OpenStack community.

The other option with OpenStack is to get the software from a vendor that provides its own branded OpenStack distribution. This typically comes with added functionality, as well as support and maintenance provided by the vendor.  Additional distribution “packaging” features make the software easier to install and operate. This reduces the need to hire additional staff that you will likely need for the DIY option as most vendors offering OpenStack distributions are also able to provide operations support and professional services. While a proprietary OpenStack solution comes with a license fee, these are likely to be much less than those associated with proprietary cloud software. The benefits of reduced staffing, reduced deployment time, and even of “one-throat-to-choke” support and maintenance may more than offset the savings that come from not paying license fees under the DIY option.

How do you choose between proprietary and open source

To summarize, there are cost and operational trade-offs that you need to consider when performing a TCO analysis comparing proprietary cloud software to an open source OpenStack solution. If you already have the staff needed to take the DIY approach, OpenStack can be a less expensive option that also gives maximum flexibility. You could even start with a DIY OpenStack approach and move to a branded OpenStack distribution later on, if merited. However, most enterprises have shown a preference for branded solutions–whether a proprietary or an OpenStack option. You can expect lower software license fees and TCO from an open source solution than from proprietary software–sometimes significantly so. Open source solutions also give you the flexibility that comes from avoiding vendor lock-in.

In the next part to this series, I will explore the issue of manpower and personnel costs on TCO. I will review how geographic location can affect personnel costs, how changes in personnel costs can affect TCO, along with the effect utilization rates have on the personnel component of cost per VM.

Other recommended resources:

For more information and resources on how to create a quick comparative multi-cloud TCO, explore the 451 Research paper, How to Create a Quick Comparative Multi-Cloud TCO Analysis Spanning Public, Private and Managed Cloud.

Read the next blog in the Cloud Economics 101 series: The Influence of Personnel Costs on TCO for Private vs. Public Clouds

About the Author


I am the Chief Operating Officer and Vice President of the Hewlett Packard Enterprise cloud business unit, driving all aspects of operations and performance. I am a leader in HPE’s Cloud Economics campaign. I have also held various leadership roles in General Electric and Electronic Data Systems, and have a Master’s degree in Business Administration, Analytical Finance and Strategic Marketing from the Indian School of Business, Hyderabad in India. I am also Six Sigma Black Belt certified. Follow me on Twitter @lalitsingh17