Breaking Through The Storage I/O Barrier For Cloud Computing
Click here to download now
Overview: It's no secret that public and private cloud computing is experiencing exponential growth. IDC projects that by 2014, server sales will grow to $718M for public clouds and $11.8 billion for private clouds. IDC also projects public cloud services will grow at 30% per year, five times that of global IT spending in 2011.
One of the biggest challenges in cloud computing is the storage I/O bottlenecks that come with large-scale virtual machine (VM) deployments. As the number of VMs grows, the corresponding explosion in random read-and-write I/O traffic brings NAS/SAN arrays or local DAS to their knees.
For public or hosted clouds, current growth is driven by e-commerce, SMBs and distributed enterprises outsourcing their platform infrastructure for cost-efficiency and resource elasticity. The result is a proliferation of virtualized multi-tenant OLTP databases and web servers that create a massive amount of random I/O traffic.
For private or enterprise clouds, Fortune 1000 data centers are struggling with spiraling VM scalability pains driven by server consolidation and emerging IT initiatives that accelerate VM density, such as VDI applications. To work around these pains points, storage administrators often resort to buying more storage array hardware, a solution that does not scale effectively in purchase cost, administration overhead or maintenance expenses (power, cooling, floor space).
This paper identifies the rapidly worsening I/O bottlenecks brought on by the explosive growth of cloud computing and how the Marvell DragonFlyTM Virtual Storage Accelerator (VSA) is uniquely positioned to solve this problem.