Rethinking mainframe storage: a hybrid cloud strategy for the best of both worlds
For the better part of three decades, hierarchical storage management (HSM) has been a key component of mainframe information lifecycle management (ILM) – and it’s been a very effective method for managing enterprise data. However, current data growth rates and shrinking budgets are challenging that structure and are forcing us to look at new, innovative, ways to manage and store data in accordance with the business value of the data and regulatory constraints governing data retention.
The need to collect information and data is outstripping storage budgets. Data is emerging as the world’s newest resource for competitive advantage, fueled by increasing collection from mobile and social initiatives. As the amount (and value) of data continues to grow, current systems won’t keep pace. According to Cisco, global data center traffic will nearly triple from 2013 to 2018 with a CAGR of 23 percent, growing from 3.1 zettabytes per year in 2013 to 8.6 zettabytes per year in 2018.
This poses tremendous challenges to HSM and ILM as we know them now. What data can I delete and when? What must I keep and for how long? What are the regulatory requirements dictating long-term retention for certain data? How can the business achieve its backup, recovery, and accessibility objectives for critical data? How do we do all of this in the most cost effective way?
One solution worth exploring is hybrid cloud storage – the combination of private and public cloud storage capabilities that often combine on-site enterprise storage with public cloud storage capabilities offered by Amazon Web Services, IBM SoftLayer, Microsoft Azure, Google Cloud, and others.
The primary benefits of this approach relate to price and economy of scale. By adding public cloud storage resources users pay only for the capacity they use (OPEX vs CAPEX), and capacity is flexible and can increase or decrease on demand. A virtualized architecture increases the breadth of available storage options, and it’s fungible. It is no longer necessary to be concerned about the type of storage or the provider! This is also commodity storage, which has been proven in enterprise settings, and its capacity is virtually limitless.
So the direction we’re advocating at Rocket Software is to use hybrid cloud as another tier for migrating to lower and lower cost storage as the retention time increases, and the value of the data decreases. This provides more flexibility, while reducing costs even further, and is respectful of current policy. We recommend moving long-term archival of aged data to the cloud, as this reduces costs and the impacts of managing data by other systems.