Cloud storage is perhaps the most critical cloud computing asset for prospective cloud providers. It may also be the most problematic, especially
Most enterprise IT spending supports core business applications that are storage-intensive, and fully cloudsourcing these applications would demand cloudsourcing the storage. But current cloud providers price storage substantially above the threshold for making these applications cost-effective in the cloud. The cloud provider must therefore either forgo cloudsourcing mission-critical applications or solve the problem of cloud storage pricing and profit.
The most compelling value of cloud computing for buyers is that in terms of total cost of ownership (TCO), it is less expensive than internally hosting applications. Price-based value propositions are always difficult for the seller, which must balance the buyer's goal of paying a lower price with the reality that the seller must have a respectable profit margin to create its own return on infrastructure investment.
Economies of scale can solve cloud storage pricing issues
The classic solution to the cloud storage pricing/profit problem is economy of scale. If cloud providers can create large resource pools and can manage those resources very efficiently, they can lower the unit cost of computing well below what businesses can attain internally. This is particularly attractive to small- and medium-sized businesses (SMBs) that can neither create large data centers nor easily acquire skilled professional support personnel.
The problem is that storage economies of scale are much harder to generate. Server and software costs are a mixture of capital cost and operations cost, and nearly all IT users say both of these costs scale in proportion to the number of devices or software installations. But storage has a very limited per-device operations cost, and storage devices are increasingly inexpensive. Users can buy terabytes of storage for a few hundred dollars -- not much more than a cloud provider could expect to pay for that volume.
Thus, the first step for cloud providers looking to ensure profit on cloud storage is to maximize economy of scale by pooling cloud storage resources efficiently. That means deploying large storage area networks (SANs) and cloud storage devices with the lowest unit cost of capacity. It's an important step, but because enterprises typically achieve good economies of scale in their own data centers, it probably won't make cloud storage profitable by itself.
Finding cost-effective methods of offering sensitive data cloudsourcing
Cloudsourcing mission-critical data also poses a threat to the network connection between the user and the cloud. If block storage or file system storage commands are sent between cloud and user, entire databases might have to flow over the interface to satisfy a simple query, creating massive congestion. But unless the user plans to cloudsource everything in the corporate data center -- which, according to user surveys, few users expect to do -- some storage data transfer between the cloud and the enterprise's on-premises resources would seem inevitable. The combined costs of the data exchange between user and cloud and the cost of the cloud storage itself could destroy cloud economics and profits alike for providers.
The most popular strategy for cloud providers addressing this cloud storage pricing/profit dilemma is to avoid cloudsourcing storage-intensive applications, but as we've noted above, that dodges the lion's share of IT spending. There are three other strategies available:
- Accept low margins, or no margins, on storage and making up the profits on compute services or other services;
- Create additional value-added services around cloud storage to improve overall storage margins; or
- Provide tools that link cloud applications to databases hosted in customer data centers, avoiding the cloudsourcing of core application storage facilities.
Some cloud operators, particularly those who are also network operators, find the first strategy at least potentially suitable for them. With historically low ROI expectations, telcos and other ISPs may have an opportunity to earn at least a small profit from cloud storage hosting -- enough that the overall hosting of mission-critical applications can be profitable.
However, operators of all types fear that setting cloud storage pricing at effectively below-market rates might attract only the high-volume users whose application hosting needs wouldn't necessarily make up for the low profits on the storage side. A price-curve approach that sets lower prices for low volumes of cloud storage may be the solution; this would make storage costs less onerous to SMBs, which typically consume low volumes of storage. Another approach is to offer lower cloud storage pricing in Software as a Service (SaaS) applications, which command higher margins.
The second option, the value-add approach, is generally seen as more broadly effective. The most attractive value-add for users is the ability to access the same storage resources from anywhere inside the cloud at the same cost and with similar performance. That can accommodate cloud applications where the preferred hosting site changes with time zone, for example. More value can be added if the storage value-add includes enabling multiple cloud applications to access the same storage resources. Finally, adding services like automatic backup and recovery, and perhaps even local data mining tools and applications, can build the value of cloud storage and increase buyer price tolerance.
The final cloud storage pricing option is the most complex to evaluate because it's essentially a way of dodging the problem. Many of today's applications are designed -- or have been rewritten -- to consume Database as a Service (DBaaS), meaning that the applications send SQL queries to a database server or appliance. If cloud applications use DBaaS to link back to the enterprise data center for access to core application data, then that data need not be migrated to the cloud. The user's data stays where it is, but the cloud provides elastic application capacity that uses the data in place. No cloud storage costs are incurred and the user retains security/compliance control.
Most cloud providers that want to develop a storage solution will likely want to adopt all three approaches. An excellent starting point is to deploy a virtual database management system appliance service in the cloud, basing the service in an established SQL-query format. This would encourage cloudsourcing applications that need both elastic capacity and access to core application data. At the same time, it provides a way to reduce traffic that might otherwise congest the networks of both the user and cloud provider, raising cost and performance risks. With proper cloud storage pricing and promotion, it could encourage cloudsourcing and manage cloud storage ROI most effectively.
About the author: Tom Nolle is president of CIMI Corporation, a strategic consulting firm specializing in telecommunications and data communications since 1982. He is the publisher of Netwatcher, a journal addressing advanced telecommunications strategy issues. Check out his blog, Uncommon Wisdom, for the latest in communications business and technology development.
This was first published in October 2011