The DDN object is real business value from more scalable storage

According to Tom Leyden, Product Marketing Director at Data Direct Networks (DDN), the issue with storage, especially with the advent of the cloud and the exponential growth in unstructured data created, is that unstructured gets difficult and expensive to store in traditional file formats. So a new way has to be found.

That new way, Web Object Scaler (WOS) is something the company has been working on for a while, and it has recently introduced Version 3.0. The company claims WOS can deploy storage for the world’s most demanding applications by using object storage software technologies coupled with a hardware stack designed to deliver rapid on-premise cloud storage with self-service access to data.

But WOS technology is not, of itself, a cloud service. In fact, the company sees it as an alternative to expensive `pay-as-you-go’ public cloud storage services, and DDN claims its linearly-scalable, global storage platform can slash total cost of data storage by more than 50 percent.

“Object storage is it is like valet parking, compared to parking your car in a multi-storey carpark yourself and later trying to remember where you parked it,” Leyden said. “Our objective is to provide the tools that allow users to build their own equivalent of DropBox services.

“So the point to note here is that this is sold as a product, rather than a service. But, it then can be used by customers to create whatever private storage-based cloud services that meet their needs. Part of that capability, for example, is being able to create active archives.”

This is an increasingly important requirement for many businesses now that big data analytics tools are increasingly a mainstream part of the business mix. In Leyden’s view, having archive data available on disk is now becoming important.

The company claims that federated WOS technology cuts away most file system bottlenecks can deliver to the point where it can deliver more than 200x performance increases over the current peak performance of even the largest public cloud storage services. This includes object retrieval of 256 million objects per second, throughput of over 10 Tbytes per second with an object retrieval latency of less than 50 milliseconds.

The key advantage with DDN’s object storage technology is its ability to scale. And the key to this is the company’s WOS 7000 appliance. This is the hardware component and comes pre-configured for rapid deployment, and the ability to run production workloads with minimal configuration. According to Leyden, this is aimed at removing the guesswork and risk from deploying cloud storage, reducing management overhead, and providing rapid time to value from storage investments.

WOS 3.0 is built on the classic `onion’ model, with the next layer up being WOS management, which now includes latency-aware access management capabilities, de-clustered data management and self-healing capabilities. The next layer provides data protection, and the newest layer is WOS Data Intelligence, which provides a new object metadata-based Search capability.

WOS 3.0 supports parallel search of user-defined metadata across a distributed WOS cloud of as many as 8,000 WOS cluster nodes. Now, users can immediately index and contextualize Petabytes of data in milliseconds versus days.

It stores objects and metadata together, which can then be searched in parallel. It is best described by example. The classic use case is searching through millions of satellite photographs of earth, where there is a need to build up a picture of events over time at a location.

At the moment the system can only store a maximum of five values, such as GPS location, time, and date. But with this it becomes possible to extract the relevant objects needed to build a content-rich `moving’ image of change at a location, over time, which has obvious values to give to tasks ranging from surveillance through to agricultural management.

DDN is targeting the Hadoop big data analytics market, and Leyden feels it will be particularly useful in searching large volumes of semi-structured data as will be found generated by the many monitoring devices that will be the backbone of many Internet of Things applications.

Though currently limited to five key value fields, he expects the next iteration to expand on that number.

Another key component is the ability of WOS to gateway to other services. It comes with its own REST-compatible API, as well as APIs for Amazon S3, Java, C++, as well as ASG and Commvault.

“Actually, DDN is totally agnostic on this issue. If users want it, they can get it,” is how Leyden defined the company’s position on APIs.

WOS now also integrates with DDN’s GRIDScalar parallel file storage system, which is powered by the company’sStorage Fusion Architecture. This is a storage array technology to tier data from file-based workflows into a simple, scalable WOS archive.

This gives users object storage and file storage on a single scale-out platform, where data can be archived for disaster recovery or worked on via a `follow the sun’ model, with the aim of maximising data protection, organisational productivity and speeding time-to-insight.

Unlike conventional file storage solutions, WOS multi-cluster configuration enables WOS to scale to any dimension as each cluster can support up to 1 trillion objects across 256 WOS object storage servers. A cluster of 32 federated WOS clusters can be architected into a single storage environment capable of servicing over 32 trillion unique object IDs without slowing performance or adding complexity to the environment.