By Mark Peters
Do you wonder what the next “wave” of storage means for your organization? Don’t worry: It’s not really “iWikiCloudStorage 2.0,” but sometimes it might feel like that unless you know what you’re really dealing with.
Web 2.0 101
Web 2.0 is basically a catch-all phrase for a world where everything is connected to everything; content is created everywhere; physical connections become less meaningful; the value of data is hard to determine, but the volume grows rapidly and is hard to predict; and file-based storage rules a world full of unstructured data. There are companies whose whole business depends on the Internet, such as eBay, Facebook, and YouTube. Their business models all share characteristics typical of Web 2.0: They are highly collaborative and interactive, reach a broad audience, and are dominated by user-generated content.
Yet Web 2.0 is not limited to social networking sites. There are many traditional businesses that can benefit from Web 2.0 practices. At the enterprise level, internal applications such as Instant Messaging and Microsoft SharePoint enable improved communication and information sharing. Large enterprises are also building external Web 2.0 businesses to provide value-added services to their customers. While it’s obvious that virtually all Web 2.0 models require additional storage, there are also Web 2.0 businesses that are actually built on rental-based online storage provisioning, including backup services, online vaulting, and data archiving.
The number-one implication of Web 2.0 would, at least superficially, seem to be growth. However, data growth has been a given since the days of the first RAMAC. What makes Storage 2.0 different are the operational demands to provision and manage that growth.
The traditional approach to supporting applications leverages silos of over-provisioned servers, network resources, and storage systems to ensure the infrastructure remains transparent and available. Everything is designed to support the fairly predictable transactional demands of the system. This is already a challenging task, but in the Web 2.0 world, in addition to the existing challenges, everything is reactive and requires supporting the uncertain and dynamic demands of massive amounts of persistent data.
Trying to run Web 2.0 applications on Storage 1.0 systems is not feasible. In a nutshell, Web 2.0 applications require a different approach to storage.
The “shopping list” for Web 2.0 storage appears daunting. The new infrastructure must be able to dynamically adjust in any dimension, including capacity and performance, non-disruptively and in real-time. In addition, the infrastructure should be based on inexpensive commodity “bricks” with sophisticated data-management attributes (reliability, data protection, migration, and tiering).
The “bricks” should be self-healing and self-tuning, virtualized, and able to handle block- or file-based storage. Avoiding vendor lock-in is also advisable. And, of course, it has to be affordable.
The good news is that Storage 2.0 is not predicated on futuristic technologies. Most of the elements already exist or are in development. A variety of vendors have, or are developing, brick-based architectures and standards-based management and search schemas, as well as powerful migration tools that can move persistent data onto more-economical storage devices. In addition, there are plenty of methods to optimize capacity (e.g., data de-duplication, compression, thin provisioning, and virtualization) to reduce costs.
Storage 2.0 is not some blue-sky concept on the horizon; it is a collection of available tools and approaches that can be implemented today. However, relatively few organizations have taken the plunge, and very often the question comes down to the build-versus-buy decision.
Build versus buy
Companies such as Google built their own Storage 2.0 infrastructures and have developed specialized expertise in massively scalable, parallel storage systems. This is not likely to be the preferred approach for most companies. Leveraging commodity disks and building a proprietary file system take many years and huge resources. Most companies will prefer to find off-the-shelf solutions.
The requirements for Web 2.0 storage are very different from typical e-mail or database applications, and not just because they’re file-based rather than block-based. Successful Web 2.0 applications can lead to thousands or millions of users, and the data (images, audio files, and video) may need to be stored, essentially forever. Remember that your existing storage infrastructure is unlikely to have many of the attributes you need. Instead, take a look at next-generation storage systems with massive and flexible scalability and predictable and reliable performance. Additionally, being expected to store billions of files and objects, such systems must also be easy to manage across multiple systems and/or locations.
Speed, flexibility, and uncertainty are the watchwords of the Web 2.0 world. Providing the storage infrastructure to support Web 2.0 environments requires new thinking as much as it requires new devices. Just because we use terms such as commodity or utility does not mean that storage management gets any easier. Indeed, the supply (or provision) side probably gets commensurately more challenging as the demand (or consumption) side gets simpler. But that’s as it should be. Users do not care about protocols, LUNs, or maintenance windows. In terms of new thinking, be glad that not everything has to be perfect. A colleague from the disk drive world recently told me that devices are actually being “optimized down” to deliver a great Web 2.0 experience as opposed to a perfect transactional result: A few dropped bits in a Web video matter much less than the time taken to re-read or recover them, which could spoil the streaming. In the traditional world—of bank statements, for example—every bit counts.
Web 2.0 is an opportunity for almost every business. A willingness to embrace Web 2.0 is a start but, as in so many things, it is the underlying basics that are crucial and will determine market winners and losers. Storage 2.0 is a key foundational element; it’s not outlandish “iWikiCloudStorage 2.0” vaporware. The main elements do, or will soon, exist. The sooner you start to plan and implement, the better prepared you will be.
Mark Peters is an analyst with the Enterprise Strategy Group research and consulting firm (www.enterprisestrategygroup.com).