Sharad describes key aspects of a data management solution, including use of any data source, target, service level objective, location, use case, consumption model and business model. With those considerations in mind, the next question is, what are the design principles for a data management solution?
At the highest level there are two primary design principles: a Software-Defined Platform and Multi-Dimensional Appliances. Let’s unpack both.
Software Defined Platform
In order to support the wide range of capabilities required, a robust data management solution requires a significant amount of flexibility. The most efficient method of delivering that flexibility is through a software-defined, API-based platform. Below are some of the core tenets of the software-defined platform:
Form factor: While the software-defined data platform can be delivered as an integrated appliance, the same capabilities can be obtained in a software-only form factor, installed on a software-defined hardware platform on-premises or in the cloud.
Data services: Software-defined applies beyond form factor — it also pertains to the platform’s ability to provide flexible data services. A single software-defined platform provides the full suite of data protection capabilities from archive, long-term retention and backup / recovery to disaster recovery and business continuity spanning the entire RTO / RPO spectrum.
Cyber recovery: It supports the ability to recover your data on-premises or in the cloud, and the capability to recover data in the event of a ransomware attack by providing secure air-gapped solutions.
Efficiency, security and integrity: The solution should support data reduction techniques such as compression and deduplication while ensuring the safety of the data through encryption and its integrity through a data invulnerability architecture.
Data management: The software-defined platform supports a wide range of data management use cases from fast, lightweight copies for dev / test, analytics, etc. to various compliance use cases, including HIPAA, SEC and GDPR.
Flexible architecture: The solution is architected using a flexible and scalable modern, services-based platform, enabling support for a full spectrum of workloads ranging from traditional enterprise applications to modern cloud native applications.
Data Protection Evolution
Access methods: The platform supports a variety of access methods, including full data restore, application-directed recovery, and API access for third-party integrations. The API architecture enables the full power of the platform through published, stable and well-documented APIs.
Consumption methods: It provides the ability to consume capabilities either as a platform managed by the end user or as a SaaS offering, which is managed by the provider.
Automation: The platform embeds and leverages artificial intelligence and/or machine learning techniques to automate commonly executed workflows to place data on the correct tier and media type, detect and mitigate system and security issues, provide access through NLP channels, etc.
The second foundational tenet required to complement a software-defined platform is multi-dimensional appliances, and those core elements are:
Scale: The appliance must have the ability to scale in place, and to scale up and scale out, while starting with a small size and adding additional capacity either through more disks or flash drives or enabled through licensing in the same form factor. It can scale up by adding more disk or flash trays behind an existing controller. And, it can scale out by adding additional appliance capacity units.
Media: The type of storage can be traditional spinning disk media, all flash, or emerging media such as Non-Volatile Memory express (NVMe) and next-generation Storage Class Memory (SCM). A traditional backup storage scenario may leverage all HDDs, possibly complemented with small amount of flash. Alternatively, the high-performance of all-flash media may be optimal for dev/test and analytics use cases.
Deployment: The same appliance configuration can be deployed on-premises in an integrated form factor, or in a software-only form factor writing to commodity protection storage. It can also be deployed in the cloud as a software-only appliance writing to object storage or offered as SaaS by a service provider.
Use cases: The appliance is designed to support a full range of software-defined platform use cases, ranging from traditional, capacity-oriented use cases (archive, long-term retention, backup, restore) to performance-oriented use cases (replication, disaster recovery, dev/test, analytics).
Security and integrity: The appliance supports security capabilities such as encryption in place and in flight plus key management. It also supports data resilience and integrity through the data invulnerability architecture.
Management: The multi-dimensional appliance can be managed through traditional on-board system management techniques. It can also be managed through a SaaS-based management portal that can manage large, multi-site environments. Additionally, the platform provides rich APIs for third-party integrations and custom, end user workflows.
Resiliency: The appliance auto-discovers component and system failures, as well as security intrusions and anomalies. In addition to alerting the administrator, it attempts to remediate the fault through a self-healing architecture or block out suspicious activity and data sets.
High availability and non-disruptive operations: The system provides high availability and non-disruptive operations (NDO) through component-level redundancy and heuristic-based predictive software that proactively discovers, isolates and remediates failure. The system provides the ability to upgrade different software and firmware in the system non-disruptively and with minimal operator intervention
Search and Analytics: The multi-dimensional appliance provides rich search and analytics functionality. It provides predictive search capabilities at the VM level, files within the VM and even content within those files. It provides detailed analytics on the nature of the stored data from the type of files, to their age, to the sensitivity of the content.
Efficiency: Efficiency is applied in the form of data reduction techniques such as deduplication and compression, which consequently reduces bandwidth use when sent over the wire. It is also cloud aware – so, for example, when searching a data set stored in the cloud it only displays the catalog and only selectively downloads the files needed to reduce cloud egress costs.
Performance: The appliance supports a wide range of performance characteristics in support of a broad range of RPOs and change rates. It supports a sufficient number of ingest streams and ingest rates even to support zero RPO (i.e. no data loss) on a rapidly changing workload.
Hear more about Dell EMC’s perspective regarding data management in a CUBE Conversation with Beth Phalen (President, Data Protection Division) and Sharad Rastogi (SVP, Product Management) during Dell Technologies World 2019.
Beth Phalen and Sharad Rastogi share the Dell EMC perspective at theCUBE in Las Vegas
Modernizing IT should be a priority for all organizations as data continues to power the future of business. Effectively and efficiently protecting AND managing that data to drive business outcomes may determine who wins and who loses in the race toward data insights. The design principles above provide guidance toward identifying the appropriate data management solution attributes for your company.