Harnessing the “Storage Cloud”

http://www.sun.com/emrkt/innercircle/newsletter/0409/feature-itm.html| 11/23/2007 2:55:55 PM
Open storage comes to cloud computing — opening up new opportunities for forward-looking IT managers

The cloud computing buzz continues to build, and a few early-adopter IT managers are already achieving the promised benefits: increased hardware utilization rates, massive scalability, lower capex on new infrastructure, less opex on licensing, training, and administration.

But here’s the thing. There’s a lot more to cloud computing than hardware efficiency.

Cloud computing is also a programming revolution — a way to access agile, lightweight Web protocols and accelerate development cycles. It’s a way to do everything from testing to service deployment. It’s a way to access services on demand: payment services, security services, collaboration as a service, and so on.

Equally important but often overlooked: the cloud opens up new opportunities for storage. The “storage cloud,” a combination of virtualized storage infrastructure and data, can provide opaque, file-oriented data access at massive scale, hiding replication and redundancy while offering enormous elasticity.

Sun takes the benefits of the storage cloud a step farther with its open storage approach, combining open source software with industry-standard hardware. Using open storage to enable a storage cloud makes it possible to cut your reliance on high-priced hardware, take advantage of open source software, save big money, and create a new breed of data-driven services.

Are you ready to take advantage of the storage cloud? It is not too early for IT managers to begin exploring the possibilities. Here are a few suggestions — from using currently available public storage clouds to building your own private storage cloud and offering storage cloud services.

Option 1: Put some data in the cloud

The storage cloud concept is still new to many IT professionals. One way to get your feet wet with minimal cost and risk is to “cloudify” a small subset of your data; for example you could start using an online backup service rather than archive all of your data on premises. This approach, which many consumers already use to back up their personal PC-based data, can give you an immediate cost savings, freeing up space on overworked systems and possibly allowing you to delay or avoid additional storage hardware purchases.

Obviously, security concerns and service level issues are vital considerations, but the bank analogy applies here: Why keep your money in your mattress when you can keep it in a bank? A new breed of third-party service providers has emerged, specializing in the safe upkeep of data. Service-level agreements have also evolved rapidly, and service providers are introducing additional transparency mechanisms so that you can monitor their systems (and your data) in real time should you so desire.

Option 2: Build your own storage cloud

Many IT departments are extremely reluctant to entrust any portion of their data to a public storage cloud, for obvious reasons. Data drives their business; it is their competitive advantage; it includes sensitive customer information. For these companies, building a private data cloud for internal use may be a good starting point.

Private clouds are on-demand infrastructure owned by a single entity that controls which applications are run where. You own the server, network and disk; you decide which users are allowed to use the infrastructure.

This approach has the advantage of offering (and requiring) significant lock-in to corporate standards. Construction of a private cloud also gives you an opportunity to synchronize corporate standards with increasing levels of open, public data standards so that you will eventually be able to “flex” your private cloud with public cloud capacity-on-demand capabilities.

There is no single methodology for building a data cloud because business requirements, IT staff capabilities, and financial realities vary widely; however, tools and services for cloud construction are evolving rapidly.

The momentum behind cloud construction is also growing as enterprises and service providers gain experience with the cloud model. The high growth rates of cloud offerings available today will no doubt accelerate that momentum; Sun expects to see huge growth in the deployment of private clouds in the near future.

Using open storage to enable a storage cloud makes it possible to cut your reliance on high-priced hardware, take advantage of open source software, save big money, and create a new breed of data-driven services.

Option 3: Offer storage cloud services

IT departments that gain experience and expertise with storage clouds will undoubtedly create new services and capabilities that will be of interest to others. You can begin to transform your private data cloud offerings into public offerings that generate revenue; or you can lend your expertise to other departments, lines of business, or external organizations for a price.

Of particular interest right now is a new breed of analytics tools that make it possible to do data mining in the cloud.

According to Jerry Held in an article for Virtualization Journal (August 5, 2008), “the most innovative new DBMS software combined with new cloud computing services... changes the economics of business intelligence (BI) and enables a variety of new analytic data management projects and business possibilities.”

One example: Panorama Software has introduced PowerApps, a cloud computing analytical engine for BI applications. It’s essentially an “Analytics as a Service” OnLine Analytical Processing (OLAP) platform that makes it possible to develop analytical applications using the power of cloud computing. Another example is Hadoop, the free Java software framework that supports data-intensive distributed applications. And the MogileFS filesystem enables horizontal scaling of storage across any number of machines.

This could also be an excellent time to consider developing analytics tools for your particular data sets, possibly with an eye toward marketing them to other organizations down the road. Pharmaceuticals companies, for example, are beginning to apply new tools and techniques to get more information out of the data they already have, helping to cut the total cost of developing new medicines.

“By moving from a mentality that sees data ‘closed’ by default to one in which data is ‘open’ by default, we have much to gain,” blogs Paul Miller in Does the Semantic Web Matter? (January 14, 2009). “By embracing ‘the Web’ within our applications rather than continuing to see it, practically, as merely an adjunct, we can unlock more of the potential that already exists.”

The importance of open data

Whether you’re considering using a public storage cloud or building your own private cloud, open data is going to become increasingly important.

How do you write cool new Web services if the data that feeds them is inaccessible, incompatible, limited to your own company’s data stores, or prohibitively expensive to obtain?

How do you create new customer service applications if you have to worry about onerous licensing and intellectual property issues surrounding data use?

And what good is cloud computing if developers can’t get the data and tools they need without risking lock-in on a particular vendor’s platform?

What’s needed is both open data in the clouds and “data clouds” built on open standards. The open data cloud can provide both an opportunity to unlock the data that’s already coursing through corporate networks and a way to gain free, unfettered access to data from many other sources.

Equally important, since the APIs to the cloud aren’t the same as those to a filesystem or a database, you will need to make sure your applications are written to them (or can be ported to them). Open APIs are critical to ensuring that your data doesn’t end up “locked-in” the cloud. In addition, you will need to look closely at the issues of data movement — what stays in the cloud, how does it get populated, and what moves to/from the compute side.

Sun’s Open Storage model

The combination of Sun’s cloud computing solutions and Sun’s Open Storage initiative delivers new options for IT managers seeking to create and/or harness the storage cloud.

The storage cloud is about managing petascale data, and Sun’s Open Storage server and storage systems can radically improve the data-intensive computing emerging in the cloud.

Open Storage is a key enabler for building a private storage cloud. It enables the cloud to be built with less complexity, at a lower cost, and at larger scale than traditional, proprietary storage. Open Storage is about using industry standard components to create low-cost, high-capacity disk drives and combine them with enterprise-class open source software to deliver inexpensive, high-capacity and highly scalable architectures.

Open Storage also enables new architectural models for data management. With the open source storage stack running on industry standard hardware, including x86 and SPARC servers, SATA drives and solid state storage, we’re able to reduce the gap between processing and storage I/O. This simplifies data-intensive computing by eliminating the need to move data around a network.

Equally important, the public Sun Cloud Storage Service, due later this summer, will provide easy and affordable access to proven, open, flexible and scalable storage resources in the cloud — with minimal barriers to entry and exit. Sun is committed to enabling a world of many clouds — both public and private — that are open and compatible, and is leveraging its extensive portfolio of products and services, and its expertise building open communities and partner ecosystems, to make this vision a reality.

Learn more

Get more information about Sun and cloud computing and take advantage of the opportunity to review and comment on Sun's Open Cloud APIs. The latest news about Open Storage and Sun’s storage cloud offerings is available here.