Backup storage infrastructure is notoriously expensive and cumbersome, but data accessibility and retention is needed to address business intelligence, test and development, malware protection and data privacy regulation compliance requirements. Storage professionals should consider developing a primary storage protection that is more self-protecting, to enhance data protection and at the same time reduce or eliminate backup storage infrastructure requirements.
There are a few key capabilities in a self-protecting production storage infrastructure. Arguably most importantly, a self-protecting production storage system should embrace the cloud for on-demand and elastic infrastructure – but it must do so in a way that mitigates latency and egress fees associated with data transfers. Many enterprises have an on-premises edge or core data center storage implementation as a cache for their “hottest” data, and then they tier their snapshots or backup copies to a public cloud storage service. In theory, this approach accelerates performance via the “hot” on-premises storage deployment, and it unlocks a lower-cost data repository in the public cloud enabling the storage of larger volumes of data for longer periods of time, more cost effectively. However, problems occur when a cache miss occurs.
A cache miss occurs when some data that is required for processing is not available in the cache memory. It must then be located and fetched from wherever it resides. In the edge/core-public cloud implementation previously discussed, this causes significant execution delays because the data must be retrieved from the public cloud, adding substantial latency. Not to mention, it also adds egress fees, which are fees that public cloud providers charge for pulling data back on-premises.
Cache misses are a growing concern because not only are they becoming more common in today’s age of booming growth, but they are also a more significant impediment as the business relies more on data-centric applications such as databases. Storage Switzerland advises storage professionals to consider an approach whereby hottest data is still cached on premises, but also that an intermediary storage service is used as a cache for “warm” data, in between the hot and cold tiers, to accelerate application performance and to control costs in the event of a cache miss at the premium performance layer.
Our on demand webinar with ClearSky, “How to Design Self-Protecting Production Storage and Gain Backup Independence,” has more detail on preventing against cache misses and architecting a production storage infrastructure that facilitates greater data availability.