Critical.
Authoritative.
Strategic.
Subscribe to CIO Magazine »

Cloud computing: Why you can't ignore chargeback

Chargeback in the cloud computing arena is the same kind of highly granular pricing associated with use of internal resources.

One of the most controversial issues in cloud computing is chargeback: pricing consumed resources on a granular basis. Amazon posts its prices publicly for all to see -- so much for compute, so much per gigabyte sent over the network, etc. Chargeback in the cloud computing arena is the same kind of highly granular pricing associated with use of internal resources.

The reason chargeback is controversial is not, perhaps, for the reason you might expect. You might think that resource users would prefer to avoid chargeback is because they currently get resources for free, and implementing chargeback would result in more budgetary outflow.

But you'd be wrong. Resource users already pay for consuming IT resources, it's just that the costs are typically quite lumpy -- perhaps so much per month for a server (whether physical or virtual), plus an associated IT overhead cost where large swaths of IT's costs are assigned to user departments on some kind of budget algorithm.

Why Chargeback is Controversial

One reason chargeback is controversial is that it's really hard for IT groups to accurately assess the actual cost of a resource on a granular basis. Different types of direct IT costs (e.g., servers vs. storage) get assigned to different budgets and groups within IT, which makes it hard to roll up costs to get a true picture of resource costs. This is not to mention the fact that other costs (e.g., power) may actually be paid by entire different divisions of a company (for example, data center power is often paid by the facilities organisation).

In this sense, IT chargeback reminds me of hospital billing. If you've ever tried to get an accurate accounting of the fee for a hospital service prior to getting the bill, you know that it's impossible. Hospitals just don't work that way. Every group in the hospital does its job, assigns its service codes, and at some point down the road all of the individual cost assignments are totaled up and sent along in a summary bill.

So too does IT do its job and eventually tots up its costs and divvies up the number among different user groups, but good luck if you want to find out what the real cost, say, of an individual server hour is.

Another reason chargeback is controversial is for an unexpected reason -- some user groups prefer confusion in billing because they "overconsume" IT resources and, consequently, get a sweet deal on their costs. Of course, for every "overconsumer" there's an "underconsumer" that's paying for more than they're using, but...Mr Overconsumer is going to fight chargeback tooth and nail.

There's been a lot of discussion that perhaps chargeback is not necessary -- that just sharing information with resource users is sufficient. This is referred to as "showback," with the notion that a resource user, when confronted with how much, for example, storage it's consuming, it will act to conserve resource use. In general, I feel that is a naive belief, borne out of a hope on the part of IT groups that indicating resource use will enable them to avoid the hard work of identifying specific costs for resources that are consumed.

And that's really what's at the bedrock for these discussions about chargeback -- accurate cost assignment to those who use IT resources. With Amazon Web Services, it's very clear how much one gets charged, and that the charge is based on Amazon's costs (with a profit margin for Amazon as well, of course).

Regarding internal IT resources, chargeback is designed to accomplish the same thing -- to ensure that all IT costs are associated with granular resources and also that the costs of those granular resources are accurately assigned to those who consume them. That's it. Simple. But not easy.

A Crucial Business Need

But the controversy over accurate cost assignment is, to my mind, being joined by a second reason chargeback is controversial but extremely important. And that reason, may, in fact, be more important long term.

Let me explain by starting with an anecdote.

It's no secret that the world of media-oriented consumer electronics is transforming quickly. The rise of digital formats is upending traditional media that were played in predictable ways -- think cassette tapes played on a Walkman.

Today, however, digital media is upon us with a vengeance and is making mincemeat of traditional distribution. These changes, however, also provide the opportunity for new media services (think Pandora) that are, at bottom cloud based. Another example is the rapid conversion of Netflix's business from physical DVD distribution to on-demand digital streaming. Another cloud service.

This means that manufacturers of consumer electronics have to respond and figure out how to incorporate cloud services into their products. In other words, they have to figure out how to integrate IT functionality into their products.

I was talking to someone from a consumer electronics company and he told me that design engineers at his company obsess with how to remove an additional one dollar of cost from their devices. The pressure to cut costs in these hyper-competitive markets is relentless.

So imagine that you're building an IT-infused device and need to keep your costs down. It would be pretty important that you understand exactly how much it costs for the IT portion of that product -- on a very granular basis.

And that's why chargeback is going to become even more important. It will move from accurate cost assignment (essentially a budget assignment exercise) to a product design cost element, in which knowing exactly what the total cost structure of the product is crucial.

When designing a product, input cost (i.e., resource price) is an extremely important signaling mechanism. Only when the true cost of IT resources used by a product are known can a product designer forecast whether one cloud service or another is a better decision. Absent accurate prices, the designer faces the danger of mispricing the end product.

In other words, chargeback is moving from an overhead cost assignment exercise (interesting, but perhaps not really crucial) to a cost of goods sold effort and, believe me, understanding the cost of goods of a product is vital to determine what to offer and how to set its price.

Therefore, chargeback is going to become a more important issue in the near future and pressure will build on internal IT groups to truly reflect -- with great granularity -- the fully loaded cost of delivering specific resources. The chargeback controversy will soon shift from whether it's important or not to how soon it can be implemented -- and I wouldn't want to be the CIO who responded with a "we're thinking about it."

Bernard Golden is CEO of consulting firm HyperStratus, which specialises in virtualization, cloud computing and related issues. He is also the author of "Virtualization for Dummies," the best-selling book on virtualization to date.

Follow Bernard Golden on Twitter @bernardgolden. Follow everything from CIO.com on Twitter @CIOonline

Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

More about: Amazon, Crucial, etwork, Netflix, Stratus
References show all

Comments

1

Imran Anwar

Mon 08/11/2010 - 13:25

Good points, Bernard. I also think another similar but somewhat separate concern for future budgets comes into play. Just like the hospital analogy, it is going to be far more difficult to project what compute/network/store resources a department or team will use in the immediate and near-term futures.

The current model of, "Hey, our budget is $X per year, and we add 10% every year to it," is far easier for business, tech AND finance types to deal with than adding a great deal of internal and granular predictive analytics to usage patterns to project future budgets.

Regards

Imran Anwar
IMRAN.TV

2

shelli.zargary@morevrp.com

Thu 11/11/2010 - 22:12

Hello Bernard,

Your article is right on target, and we are sure that market forces will not allow IT teams or anyone else ignore chargeback once they find out that they can get all the details they need. In fact, when cloud providers and customers realize that they can obtain the granular figures on their current cloud resource consumption to monitor, analyze and then plan future resource usage, a new chargeback based billing model will certainly gain popularity.

You mentioned the two current obstacles to chargeback: "One reason chargeback is controversial is that it's really hard for IT groups to accurately assess the actual cost of a resource on a granular basis." and "Another reason chargeback is controversial is for an unexpected reason -- some user groups prefer confusion in billing because they "overconsume" IT resources and, consequently, get a sweet deal on their costs."

We at MORE (More IT Resources) have a solution to both aspects of the problem – 1. our MoreVRP for Cloud Computing solution both accurately measures and monitors resource consumption in real-time at the highly granular specific-transaction level as well as at the tenant level; and 2. we have developed a pay-per-usage (chargeback) billing model that will enable "tiered SLA" – providers will be able to accurately bill customers per usage and will be able in tandem to offer differently priced service packages for different levels or tiers of QoS/performance. The solution is in production at some select customers and is planned for GA soon.

Shelli Zargary

Post new comment

The content of this field is kept private and will not be shown publicly.
Users posting comments agree to the CIO comments policy.
Login or register to link comments to your user profile, or you may also post a comment without being logged in.
Related Coverage
Related Whitepapers
Latest Stories
Community Comments
Tags: cloud computing, internet
Latest Blog Posts
Whitepapers
  • Oracle x86 Rack Servers Optimized for Rapid Deployments and Operational Efficiency
    Business-critical and mission-critical workloads — demanding applications and databases — require stable and secure environments. When these types of workloads are deployed on x86 servers, the need to ensure business continuity, maximum uptime, and consistent processing means that IT managers and business unit managers are looking at enterprise x86 servers in a new way: They realize that the business depends on these servers and that x86 server platforms for the enterprise are no longer expendable, as they might have been when servers were dedicated to a single application — or when they were deployed as small Web servers that could be easily taken offline and replaced.
    Learn more »
  • Delivering Tomorrow's Backup and Recovery Infrastructure
    The data protection market has changed considerably over the past decade. During this time, the market witnessed a fundamental shift away from relying solely on tape for backup and recovery to using disk-based backup solutions to address challenges including backup performance, reliability, and recovery time objectives. This paper highlights that firms evaluating next-generation data protection solutions must expect a greater degree of integration between the technology components in today's data protection path.
    Learn more »
  • 2-Layer BPM: Oracle's Unique Strategy Towards Exceptional Agility and Business Process Efficiencies
    Today, a new approach to BPM — the use of BPM and SOA together in a layering strategy — offers built-in smartness and high configurability. This dynamic approach to business process management is based on context and content. It offers agility throughout an organization, and it can dramatically increase productivity and time-to-market.
    Learn more »
All whitepapers
rhs_login_lockGet exclusive access to Invitation only events CIO, reports & analysis.
Recent comments