- Open Data Center Alliance Usage Models
What is an Open Data Center Alliance Usage Model? The eight Open Data Center Usage Models address the most critical requirements facing IT today in the adoption of cloud computing as well as the Open Data Center Vision for Cloud Computing. The documents are user-driven cloud usage models and data center requirements.
What is the purpose of these Usage Models?
Alliance publications document a unified customer voice on requirements for widespread adoption of cloud and next-generation data center solutions. Open Data Center Usage Models prioritize and document the most pressing requirements customers face today, reflect input from over 300 IT leaders, and are expected to help shape industry solution delivery and accelerate adoption of cloud services based on open, industry-standard, and multi-vendor solutions.
Who created these documents?
The documents were created by Alliance members. Specifically, Alliance Workgroups – infrastructure, management, regulation & ecosystem, security and services – led by a Technical Coordination Committee to prioritize usage model targets based on broad member input.
What are the published Usage Models?
Usage Models include: Cloud Provider Security Assurance, Security Monitoring, Service Catalog, Standard Units of Measurement, Virtualization Interoperability, I/O Control, Usage Regulation, and Carbon Footprint.
- Cloud Provider Security Assurance & Security Monitoring address IT's largest challenge for cloud adoption by proposing standard security levels for cloud services & compliance.
- Service Catalog & Standard Units of Measure enable feature, price and performance comparisons across private and public clouds for increased transparency and easier IT decision making.
- VM Interoperability & I/O Control addresses the technical foundation required for federated cloud interoperability & improved quality of service.
- Regulation & Carbon Footprint outline expectations for cloud services to ensure compliance to government and corporate reporting requirements and outlines a means for services to be CO2 aware for subscribers.
What has been the response to the usage models? The Alliance stated in June, 2011 that it expected initial industry response from solutions providers and service providers within 6 months of publication. In September, initial solution provider member response was exhibited in a series of usage model proof of concept demonstrations at the Intel Developer Forum event. Demonstrations from members including Cisco, Citrix, Dell, EMC, Joulex, Parallels, Red Hat and VMware highlighted usage models including Carbon Footprint, IO Control, Security Monitoring, Service Catalog, and VM Interoperability. We expect this early industry delivery will help shape product delivery meeting usage model requirements in the months ahead and signals a broad industry focus on addressing Alliance usage models.
What does the Provider Security Assurance usage model address?
Security is arguably the biggest challenge customers face today as they have no way to understand or specify the specific security requirements of a particular cloud vendor or if cloud service security levels violate regulatory and internal security policies. The usage model documents requirements for a granular specification of security capabilities delivered by service providers. It accomplishes this through a tiered model for differentiation of service delivery enabling competitive offerings which allow for feature and cost differentiation. This allows enterprises a standard way of determining what security capability is being offered.
What is the need for a Security Monitoring usage model? Once you have established a standard framework for security capabilities, you also need to be able to determine that service providers are meeting the levels promised. The usage model describes the need for specific hooks required to enable real-time monitoring of cloud service security, defines what needs to be monitored, and documents a need to be able to deliver monitoring to both customers and third party certification agents.
Why was the Service Catalog usage model prioritized? The service catalog addresses today's fundamental issue on how to measure what services are being delivered and what attributes exist for the service. Today, there is no standard, apples-to-apples comparison in the marketplace, making provider selection both timely and difficult. The Service Catalog defines the need for specific definitions on what is listed in a catalog framework as well as the standard attribute characteristics.
What is a Standard Units of Measurement usage model? Today's metrics are very granular, but end customers view cloud services at a much more aggregate level. The gap today is measuring services based on performance, availability, security, reliability (QoS) etc., with metrics designed for cloud computing models.The usage model defines what framework and attributes candidates for metrics should have.
What are the issues addressed in the VM Interoperability usage model? Users would like to select any cloud vendor based on cost and performance/capability characteristics. They would also like to link their private clouds to public clouds – either for bursting of specific workloads or for true federation of cloud services. Today every cloud provider chooses a particular implementation of virtualization whether it be one or multiple hypervisors. While today OVF (Open Virtual Format) provides migration of workloads at a container level allowing for migration of VMs from heterogeneous hypervisors, there is no way to consistently manage VMs in a heterogeneous hypervisor environment. The usage model specifies which areas associated with VM interoperability need to be resolved: creation of new management interfaces that are consistent across all hypervisors to work with OVF to enable true interoperability of hypervisors.
What issue drove the prioritization of the I/O Control usage model? Today's virtualization solutions offer no clear method for establishing I/O QoS levels for individual virtual machines allowing no standard mechanism to manage IT SLAs. Today, this results in the potential for unbalanced I/O across VMs and risking instances of single apps or users consuming full platform bandwidth. The net result is often an artificial lowering of VMs per physical server based on I/O bottlenecks causing inefficient use of physical systems. There is a key need to control bandwidth allocated to each app and a quota of bytes applied to each VM. The Alliance's I/O Control usage model outlines requirements for hardware and software mechanism delivery to allow data center managers to specify exact bandwidth per VM and total amount of network I/O per user based on policies.
Why is a regulatory framework needed? Not all cloud adoption issues are technology related. Many are related to government requirements for IT as well as policies established by vendors that impede the adoption of cloud. A holistic approach to use of the cloud weighing both technology attributes and policy are equally important, and a unified view of regulations will help create more efficient engagement with regulatory agencies, potentially saving enterprises significant investment in the unilateral engagement pursued today. The Alliance's regulation document providesacompilation of government agencies and regulatory bodies from around the globe that are engaged in topics associated with cloud computing.
Why is a Carbon Footprint usage model important? Many corporations are providing carbon reporting of IT as part overall sustainability and CSR initiatives as well as emerging government requirements. IT must be able to measure cloud service carbon footprint as part of these larger corporate reporting requirements. In addition, many IT organizations are prioritizing lower carbon emissions through efficiency initiatives, and carbon footprint reporting would assist identification of green service providers.The usage model addresses this by defining a requirement of workload execution level carbon reporting as part of customer reporting to enterprise.
| 
