Copy Data Corner Chapter 2: Data On-Demand
The Goal of IT as a Service
Speak with most IT leaders about their big picture strategies and goals these days and most will describe a cloud-like delivery model in which IT outcomes are provided rapidly, efficiently, and in many cases, via self-service. And these same leaders increasingly are looking to leverage multiple clouds to augment and/or replace existing application infrastructure. In this new world, IT delivers “services” across the stack – accountable to service-level objectives (SLO) or agreements (SLA).
This service-management mindset is hardly a new concept – ITIL has been a process focus/discipline for decades. Rather, the maturity and automation-capabilities of the underlying tools and technologies finally caught up to the premise. Compute, network, and storage virtualization enabled a more automated infrastructure stack. SaaS offerings that automated business processes like CRM, ERP, HCM offloaded the off-the-shelf application management burden. And service management platforms like ServiceNow provided a means to automate across an entire range of IT services.
In spite of this progress, many IT leaders recognize a missing link: the data that powers business applications. Unfortunately, data has grown to a size within the enterprise wherein it’s ‘gravity’ creates access, mobility, and usage barriers. Data growth rates continue to accelerate, which means the problem only compounds over time. And the volume of data consumers within the enterprise is growing just as fast. It seems that every department wants rapid – and frequent! – access to the most current data, for uses as diverse as development and test, to analytics, to training and enablement.
It’s All About the Apps
When it comes to harnessing the power of data on demand in the enterprise, one critical success factor is application-awareness. Application data that remains in its native format, in an application-consistent form, is instantly usable for any purpose. Handing developers a read/write instance of a SQL/Oracle database, from yesterday at noon, with logs rolled to the appropriate point-in-time, is instantly useful for development, test, reporting, and more. On the other hand, translation from a media format – for example tape-streaming – or reassembly of data from unintelligent database dumps or dedup backups adds time, complexity, manual intervention, and in most cases additional infrastructure resources.
True Copy Data Management (CDM) is application-centric in its approach to data capture. Point in time images are taken in the application’s native format and preserved in that format throughout the data lifecycle. Those images are captured using the most efficient, and application vendor-specific, manner possible. After first capture, only incremental changes are recorded. Yet with each incremental capture, a full, virtual point-in-time image is generated to ensure the timeliest access. The frequency of capture, application-specific characteristics, location for management, and lifecycle of the data is all managed via an SLA.
Organizations who take advantage of Actifio’s CDM software recognize that this application-centric method of data management allows them to fill-in the missing link to their IT-as-a-Service strategy. CDM allows them to deliver the missing Data-as-a-Service layer in their IT stack.
Find. Rewind. Use. Instantly.
In addition to data size and growth, one of the other challenges facing enterprises today is the proliferation of data – particularly the non-production copies, which can represent 10-20X the volume of production – and the myriad locations that data might be found. Think about a typical web application built using an Oracle database. In addition to the production instance of this application, copies likely exist in development environments (at least one), performance test, integration test, training, and pre-production. Some of these environments might be on-premises, others in the cloud. Then, add operational copies for backups and DR, wherein the same on-premises vs. cloud phenomenon exists (or mountain bunker, in the case of backup tapes). Then, add extra copies for warehousing and reporting. And copies against which to run machine learning algorithms. And copies to respond to regulatory compliance mandates.
The reality is: it’s increasingly difficult to understand where all the data copies are, let alone effectively manage them. That represents a cost and efficiency challenge. But, in today’s world, in which data is often under attack from internal and external threats, this data sprawl also represents substantial risk to the business. Every copy of data in existence increases the threat “surface area”. The rule of large numbers takes hold – the more copies, the harder it is to ensure proper security, privacy, and availability controls for all of them.
External regulations add additional elements of risk when it comes to this vast data landscape. In addition to cyber-security and availability, the privacy of consumer data is being scrutinized through directives like GDPR, from the European Union. Penalties for non-compliance can be significant.
Enterprises, more than ever before, are leveraging public cloud for some or majority of their applications. The copies of application data multiply like rabbits in public cloud as well. A few examples are EBS snapshots to S3 in Amazon AWS, database dumps to EBS, backup software to store data for many years in S3, host based replication software to replicate to a different cloud region. The infrastructure costs in public cloud add up very quickly.
Copy Data Management software directly addresses this issue of unmanaged data sprawl on-premises and in public cloud. By capturing enterprise data, and maintaining throughout its lifecycle, CDM consolidates separate data silos. CDM also allows for the search and access of any data, from each captured point-in-time, across any location. Those locations include public and private cloud, as well as remote data centers. And, since the data is captured in the application’s native format, once data is found through a simple, hyperscale search engine, instant access and use of that data on demand is always available.
In our next Chapter of the Copy Data Corner, I’ll dig deeper into the requirements of the application data economy and the evolution of operations in IT.
Get Under the Actifio Hood