Matthew Hurford, CTO South Pacific, NetApp
Data&StorageAsean : In your own words - What is Object Storage?
Matthew: Object Storage is a technology whose time has come. Exponential increases in the volumes of data being stored, the growing importance of metadata, the need for global access, and a transition to distributed cloudbased applications have pushed traditional block- and file-based storage beyond the breaking point. Object storage is not just about adding new storage functionality. It also fundamentally reduces the complexity of how applications interact with storage. The basic object model of HTTP (create, retrieve, update, and delete) has been well proven in the world’s most scalable and successful application, the World Wide Web, and the standardization of object storage follows the same proven model — dramatically simplifying how applications interface with storage. Even more exciting is how object storage allows applications to express their needs along with each stored object. This enables storage systems to provide functionality that is directly relevant to the application, and most importantly, to provide storage services that are cost appropriate for the data being stored. Aligning the costs of stored data to the value of stored data is critical to surviving the coming data explosion. This is an exciting time in the storage industry, one of transition, innovation, and growth, and ultimately a step towards solving one of our greatest challenges — how best to store the world’s information.
Data&StorageAsean: What are the major use cases for Object Storage Solutions?
Matthew: According to IDC the use case for Object storage are nay and varied. Their latest survey “Understanding End-User Object-based Storage and Trends” lists -
Traditional IT is replacing traditional file servers
Web streaming services
In general you can categorise these applications into three main areas:
Web Data repositories – Billions of small objects normally with a high transaction load using native web protocols
Data Archives – Larger objects with longer retention periods like backups, health records or e-libraries
Media Archives - Large objects that are potential never deleted ie movies that need high bandwidth for global streaming. Cisco estimate that 75% of mobile phone traffic will be video streaming by 2020.
Data&StorageAsean : Should object storage be on premise or in the cloud?
Matthew: Object storage because of its architecture is suited to either or both in a dynamic hybrid IT model. You should have object storage close to where the data is being created or ingested and close to where it is being consumed. By close we mean in terms of latency. IoT for instance can send data to cloud repository because latency is not normally an issue. Whereas media companies need the lowest latency to transform the raw media files into production quality. So they need on-premise storage during the workflow but can be in the cloud for distribution.
Data&StorageAsean : What are some of the key features people should look for when choosing an object storage solution?
Matthew: Flexibility, Software defined, maturity, geographically aware, security and intelligent data management.
Flexibility in terms deployment options. Different sized sites based on need, some software and some appliance based. Flexibility in terms of data protection polices based on latency (full copies versus erasure coding), site protection (how many sites can fail and you still have access to your data availability and durability). There are a lots of new object stores in the market, you need to pick one that has a mature product and a track record of deployments, large and small. Object stores due to their global deployments, need to be able to understand how to move data based on intelligent policies but also where not to move data based on regulation and compliance requirements. Encryption down to an individual object level based on system managed intelligent policies is a must to protect privacy and IP of individuals, companies and governments.
Data&StorageAsean : Is there anything unique about your company’s object storage offering(s)?
Matthew: NetApp delivers the following three key capabilities for its object store called StorageGRID WebScale:
A single, large repository for unstructured data. StorageGRID can store billions of datasets, such as files or objects, and petabytes of capacity in a single, multisite namespace.
Global data creation and management. Using a sophisticated global policy engine, data can be created, managed, or used at any location, while making sure that data resides where it should, as long as it should, and on the optimal storage tier.
Intelligent data classification and access. Using metadata that describes the data stored in the repository, policies can be executed in a highly granular yet efficient way. In addition, data can be managed and retrieved simply by using metadata, such as account numbers or the name of a customer, regardless of where the data is physically stored.
StorageGRID makes it possible for complex storage networks involving multiple applications using multiple protocols spread across multiple sites to all be seamlessly managed as a single entity. StorageGRID can provide secure public or private cloud storage services to multiple tenants, each with their own policies and administrators. It also allows for storage to be organized into arbitrary storage pools that can overlap and be grouped by tier.