The Growing Pains of Cloud Storage

C loud storage is growing at a phenomenal rate, fueled by multiple forces, including mobile devices, social networks, and big data. Content is created anytime and anywhere on billions of smartphones and tablets; high-resolution photos and videos are frequently uploaded to the cloud automatically as soon as they're captured. A Gartner report predicts that consumer digital storage will grow to 4.1 zettabytes in 2016, with 36 percent of this storage in the cloud. 1 Social interactions and transactions on the Internet are frequently captured and analyzed for targeted advertising. In addition to social networks and e-commerce, big data analytics are growing in many other sectors , including government, healthcare, media, and education. An IDC forecast suggests that big data storage is growing at a compound annual growth rate of 53 percent from 2011 to 2016. 2 The growth in cloud storage has made it an expensive cost component for many cloud services and today's cloud infrastructure. Whereas raw storage is cheap, the performance, availability , and data durability requirements of cloud storage frequently dictate sophisticated, multitier, geo-distributed solutions. Amazon Simple Storage Service (S3) offers 11 nines of data durability (99.999999999 percent), but some other services demand even more stringent requirements due to the sheer number of objects being stored in the cloud (1.3 billion Facebook users, uploading 350 million photos each day) and to the data's importance (who can afford to lose a video of their baby's first steps?). Data is frequently replicated or mirrored in multiple datacenters to avoid catastrophic loss, but copying it across datacenters is expensive. The networking cost is frequently proportional to the distance and bandwidth requirements between datacenter sites. Traditional storage systems use dedicated hardware and networking to guarantee preservation of the quality-of-service (QoS) requirements, such as throughput, latency, and IOPS (total number of input/output operations per second). Unfortunately, these dedicated resources are frequently underutilized. Cloud computing promises efficient resource utilization by allowing multiple tenants to share the underlying networking, computing, and storage infrastructure. However, providing end-to-end storage QoS guarantees to individual tenants is difficult without mechanisms for avoiding interference. Typically, in a cloud environment such as Openstack, multiple tenants share the backend block storage (Linux's logical volume manager or a Ceph RADOS block device [RBD], for example) through a storage virtualization layer such as Cinder, which attaches virtual machines (VMs) to individual storage volumes. Providing custom-ized storage QoS to meet different tenant needs is …