This is part three of the three-part series on working with Infrastructure as Service from Jo Maitland, Executive
Editor of SearchCloudComputing.com.
Even if the cloud's control and security issues were suddenly resolved, the morass of cloud pricing would still be a challenge. Although cloud providers like to tout the simplicity of their services, IT managers have found that pricing cloud services is anything but simple.
Pricing structures are based on a multitude of factors, from storage space needed to clock cycles used to monthly traffic allotments, and that's not all. Some service providers have additional charges hidden deep within their service-level agreements (SLAs).
To determine total pricing for a cloud service, users need to understand the individual service elements that a provider bills for and how those charges are calculated. Does the provider, for example, bill based on traffic, storage space needed, server CPU time or a combination of these factors along with other elements?
Costs stem from services
Another critical factor in determining true costs comes down to the type of service needed. For some, that service may be little more than a hosted, dedicated server to run applications in the cloud. For others, the service may be cloud-based backup or business continuity or basic hosted storage.
Perhaps, the easiest way to break down pricing is to focus on the primary services offered. The majority of cloud service providers break down their services into three primary areas: servers in the cloud, storage in the cloud, and sites and applications in the cloud. Each is governed by its own formula for pricing.
Servers in the cloud come in two forms: virtual and physical. In other words, you can purchase time on a virtual server (where the physical hardware may be shared with others) or time on a dedicated server (where you are the only "tenant" on that server). Table 1 shows the pricing breakdowns:
The pricing comparisons are only part of the overall picture; each vendor listed above includes extra services and features for an additional charge. What's more, prices can change depending on length of commitment, total bandwidth needs or total size of storage required, and in most cases, prices and packages are negotiable with a vendor's sales staff.
Different providers, different services provided
Not all cloud service providers are created equal, and that becomes evident when you look more closely at the differences between providers and how they address customer needs. For the purposes of comparison, we picked three of the well-known cloud infrastructure providers:
- GoGrid includes load balancing at no extra charge on some of its server offerings and also includes the first 20 GB of storage for free.
- Rackspace uses a different billing mechanism and reduces per-gigabyte bandwidth charges as volume increases; the company also offers backup services at no charge on some of its virtual server bundles.
- Amazon offers sliding scales for rates on most of its services but charges for gets and puts to its storage services.
When you look at the simpler concept of cloud-based storage services, the differences in pricing structures and in how providers build bundles for dent (see Table 2). Once again, the pricing comparisons are only part of the picture. And like other cloud services, prices are affected by length of commitment, total bandwidth needs or total size of storage required. And as with other cloud services, there is plenty of room for negotiation.
Once you take a closer look, the differences between service providers take shape:
- GoGrid offers the first 20 GB of storage at no charge to customers using its hosted server services and offers its storage services only to hosted server customers. Typical of the market segment, GoGrid offers discounted per-gigabyte pricing as customers buy larger storage amounts. GoGrid's storage service is offered as a mountable volume and does not yet offer a Web services application programming interface for gets and puts and other commands.
- Rackspace tries to keep its storage services billing model as simple as possible. The company offers a sliding scale for storage purchases, where the price per-gigabyte decreases as total amounts ordered increases. What's more, if the file is more than 250 KB in size, the company doesn't charge for gets and puts.
- Amazon S3 does not charge for deletes and offers discounts as transaction volume and storage needs increase. The company also offers contract-based fixed pricing for those looking to stabilize their prices. For larger file transfers, the company recommends using its import and export services, which can reduce costs.
Ideally, pricing out cloud-based services should take little more than selecting features and determining storage and server computing needed. In reality, most IT managers will find that ideals don't apply here and will have to carefully consider the "hidden extras" as well as the standard charges to determine the cost for services.
The bigger challenge is converting the "technospeak" of cloud services pricing into something that those holding the purse strings can understand and to bring those individuals into the loop to negotiate pricing and contract terms as well as SLAs.
The trick is to get all estimates in writing with clear, concise language that explains what the total charges will be per contracted period: Written confirmation is the only way to see an apples-to-apples comparison of pricing models and determine true budget amounts.
The future of IaaS
A recent study by the University of New South Wales in Australia, in collaboration with researchers at NICTA (National ICT Australia) and the Smart Services Cooperative Research Centre, spent seven months stress testing Amazon's EC2, Google's App Engine and Microsoft's Azure cloud computing services.
The analysis simulated 2,000 concurrent users connecting to services from each of the three providers, with researchers measuring response times and other performance metrics. The results were less than stellar, with response times varying widely depending on the time of day in which services were accessed. In addition, the study revealed a lack of monitoring tools to enable organizations to check whether the service has met their SLA.
It's obvious that cloud providers have their work cut out in terms of simplifying pricing models, beefing up security and providing SLAs that guarantee better reliability. The market is evolving fast, and today's dominant players could be history in two years. Just one catastrophic data privacy breach could extinguish a brand and potentially slow the market down for a decade.
Hopefully, cloud providers are listening, and their customers will be willing to keep pushing for what they need.
TABLE OF CONTENTS
• Infrastructure as a Service: How to maintain control
• How one growing firm uses Amazon's EC2
• Cloud computing concerns slowing widespread adoption
• Understanding cloud computing pricing
ABOUT THE AUTHOR:
Jo Maitland is an executive editor in the Data Center and Virtualization media group at TechTarget. This article features additional reporting by Christina Torode, a senior news writer in the CIO media group and TechTarget.
Editors' note: This chapter on Infrastructure as a Service is the fourth part of an e-book on cloud computing that also includes chapters on CIO strategies for the cloud, development for the cloud and Software as a Service.