What is the Cloud Data Management Interface (CDMI), and what does it have to do with data storage? ESG senior analyst...
Terri McClure delves into the technology.
The Cloud Data Management Interface (CDMI) is the centerpiece of an effort the Storage Networking Industry Association (SNIA) defines as “the functional interface that applications will use to create, retrieve, update, and delete data elements from the Cloud.”
An interface such as this is the key to standardizing how we’ll create and manage cloud storage data objects. It also holds huge potential for cloud storage.
Currently, most cloud storage service providers use a proprietary object storage interface. These proprietary interfaces are similar -- they’re all RESTful, with basic “put,” “get” and “delete” commands -- but they each use different semantics. It’s like the French spoken in Paris and the French you hear in Quebec. Many words and phrases are identical, but enough differences exist to make communication difficult, unreliable and sometimes even impossible.
Making multilingual systems
Cloud service providers and object storage software vendors publish object storage APIs that enable users to code applications to manage data objects. But most IT applications use block and file storage formats, not object formats. IT teams are loathe to spend money to rewrite their application interfaces to “speak object” so they can use software or cloud services, never mind spend the money to test and maintain those custom integrations.
Key CDMI functions
• Discovers the capabilities of a cloud storage offering.
• Manages containers and the data placed in them.
• Allows metadata to be set on containers and their contained data elements.
• Exposes the capabilities of the underlying storage and data services so clients understand the offering.
This situation has given rise to cloud storage gateways. A gateway system is a local virtual or physical appliance that offers IT a standards-based interface (iSCSI, CIFS or NFS) and translates standards-based storage operations into the object software or service provider’s API language. Many gateway systems do much more by offering encryption, data deduplication and compression to reduce bandwidth and transaction costs, and caching to mitigate latency associated with remote hosting.
Most of these systems are multilingual, speaking the API languages of multiple cloud providers. That way, if one vendor shuts down a service or a portion of one (as Iron Mountain did last year), IT can use gateway offerings to move data transparently to another provider. By the time Iron Mountain shuttered its service, Nasuni had already quietly migrated most of its Iron Mountain subscribers to other service providers.
CDMI = Potential
CDMI's global traction
Standards organizations around the world are adopting the Cloud Data Management Interface (CDMI) as a method of storing data in the cloud. CDMI release 1.0.2 has been submitted to the International Organization for Standardization (ISO) for standardization (as ISO/IEC FDIS 17826). Ratification should happen in 2012. The European Union (EU) is folding CDMI into initiatives, including the FP7 project from the Standards and Interoperability for eInfrastructure Implementation Initiative (Siena) and the Virtualized Storage Services Foundation for the Future Internet project (Vision Cloud) from IBM Research Haifa. Other projects/organizations, such as the OpenNebula Project, Venus-C and the National Institute of Standards and Technology/ Standards Acceleration to Jumpstart Adoption of Cloud Computing (NIST/SAJACC) are also adopting the standard.
With CDMI, problems related to porting data to other service providers shouldn’t arise. Will that be true out of the gate? I’d like to say yes. However, those of us who’ve spent many years in the storage industry bear scars from efforts like this, and we expect some bumps in the implementation road. Over time, the bumps will smooth out and we’ll see true data portability between services without having to rewrite object interfaces.
What will that development do to the cloud gateway market? It depends. “Gateway” is a misleading term as most of these systems are evolving into so much more. I refer to them as cloud-integrated or cloud-enabled storage arrays because that’s clearly the direction they’re heading. And when you’re moving to the cloud, incorporating a local-presence “bridge” from your data center to mitigate latency, ensure consistent performance, provide high availability, and offer encryption and data reduction is a necessity.
Ultimately, CDMI will make it easier and less expensive for everyone to adopt and support cloud storage. Is it “prime-time ready”? Maybe. ISO ratification of the standard will help, but the true test will be vendor adoption. NetApp is solidly behind the effort; but Cisco, Dell, EMC, Hewlett-Packard, IBM and other industry heavyweights are also involved. But in the storage industry, action speaks louder than words. The key action here involves vendors rolling out embedded CDMI functionality.
BIO: Terri McClure is a senior storage analyst at Enterprise Strategy Group, Milford, Mass.
Dig Deeper on Data management tools