Performance, Capacity and Security Key Considerations for Cloud Storage

Article by Govind Desikan



Technology advancements and business demands have largely involved the storage discussion within the cloud milieu, and 3 critical considerations for cloud storage are performance, capacity and security

Amid all the hoop-la about the cloud, storage had always taken a backseat. Cloud is viewed more as a collection of CPUs, memory, and bandwidth.

Expanding need of data is one of the key reasons for consideration of storage within the cloud. A classic example is Big Data – can I have data from across the enterprise to make business sense out of it? And, the touch points of applications used by enterprises have become wider. So customers started looking at storage for shorter time periods, for instance to run certain workloads – thus the emergence of cloud storage into the forefront.

Adhering corporate governance, compliance and regulatory requirements are forcing organizations to keep huge amounts of data that they generate for longer periods of time. Expanding storage needs are now pushing the importance of cloud storage.

Emergence of newer and disruptive technologies such as Big Data – where data can be horizontally sliced and analyzed for mining intelligence from it – required storage for shorter periods of time.

There is humongous amount of data that is being collected in enterprises today. It is generally said that there is approximately 100 Exabyte of data per year being generated – that is equal to about 50 billion iPads (32gb) being stacked together. The need for storage has hit to roof.

Now, lets talk about cloud and storage.

People invariably correlate storage to database in the earlier days – about 5-10 years back. When cloud became popular, storage within cloud was still not what CIOs were evaluating. The key reason was performance – amount of performance needed for databases was not guaranteed by cloud providers. So they started putting databases in-house and applications on the cloud.

Today, that chasm has been crossed over.

Today, the database-owned data in the enterprise is only 8% and balance 92% is unstructured data. When it comes to unstructured data the priority changed from performance to capacity.

Earlier, when capacity was the way to look at storage, vendors started looking at ways to sell more by harping on the Input-Output performance (IOPS). The issue here is that a correlation of IOPS vs capacity will never go hand-in-hand. For instance, SATA drives are now available with capacity of 1TB, but the best-in-class SATA disks can give maximum of 125 IOPS. So if an application needs 300 IOPS and 1TB of capacity, the only way to cater to this requirement is to have excess capacity.

Organizations started having excess capacities within their IT due to this trade off. This is another reason that storage discussion within the cloud came much later than compute – organizations already had excess capacities internally.

This performance-capacity anomaly may sound paradoxical, but that is a reality. The more IOPS you need, you are forced to pay more either as capacity or look at other enterprise-class storage with higher performance but also higher costs.

Performance is also linked to the latency within the cloud storage. Specially in India, organizations put database in-house and applications on the cloud. What happens here is that Internet gets added to the data access layer thereby adding on to the latency.

Performance and capacity are clearly key considerations within the cloud storage solution discussion.

When it comes to security, it is dichotomy when I say that cloud is more secure than data stored in-house within the organization.

The provider of cloud storage does not know the context of the data – which is the key to most data breaches. It is proven globally that 68% data theft is internal. For a cloud provider the data stored within his environment does not make sense to him.

Since there are so many questions on the security in the cloud, providers pump in as much security as possible. Cloud providers are always at the bleeding edge of security technologies and can replenish faster than IT departments. Within the corporate environment, the CIO is limited by factors such as thin budgets, political resistance and shortage of skilled resource.

From a technology angle, data security within the cloud – may it be multi-tenant or private – is based primarily on the securing the hypervisor. In the cloud environment, multiple virtual machines (VM) share physical logical unit numbers or LUNs, compute and memory. It is important to secure the hypervisor layer so that data is not corrupted or accessed by the wrong VM. It is important to isolate the bunch of VMs used by a particular organization so that it can eliminate unauthorized data access.

Past the hypervisor layer, security is based on the organization that stores the data – cloud provides multiple layers of security but it depends on the security and access policies followed by the organization that makes the difference.

These considerations are key to ensuring that the enterprise cloud storage caters to every business and technology need of the organization. Today cloud storage is mainstream and it will see a lot of traction from now.

At Netmagic we leverage robust infrastructure management services with innovative technologies to deliver best-in-class cloud storage solutions to clients in India.