Services-based storage, infrastructure, and data protection trends and technologies are recurring topics in this blog. Awhile back I wrote a post about enabling data protection as-a-service discussing the need for centralized management at cloud-scale, multiple service rates based on customer data protection needs or usage, and historical data for analysis and trending. The reality is that you can only get so far with legacy products built for physical environments. At some point, management tools, like the data center environments they support, need to be remade to the requirements of the day. Effective data protection solutions are no exception.
Data protection needs are more acute for as-a-service cloud models and require new approaches. Now, with the release of EMC Data Protection Advisor 6.0, I would like to share what it means to augment a successful data protection solution and extend it with a new distributed architecture and analysis engine to cloud deployments, without losing any usability benefits (i.e. without making it complex).
At Your Service Computing
There has been much discussion about the challenges of uniting disparate data center technologies. The situation has gotten so acute, however, that more radical changes are often needed compared to the more evolutionary approaches of the not too distant past. Why?
Many enterprise IT organizations are evolving to an as-a-service business model, effectively turning IT into internal service providers. This approach means extensive server virtualization, with applications supported on a changing landscape of virtual machines with loosely coupled ties to underlying storage. While appealing to application owners looking for infrastructure resources on-demand, it introduces the problem of keeping data protection policies in sync with the changing compute landscape in the data center. Case in point: when a new virtual machine is created to run a mission-critical application, the appropriate protection policy may or may not be applied, because of a manual misstep (e.g. use the wrong tool) or wrongful assumption that the virtual machine would be protected. The application may run for a while (e.g. a week) but then crash with all associated data lost. In a worst-case scenario I heard about, an outage occurred at a large financial institution where the cost of downtime is measured in thousands of dollars ($US) per minute. In this situation, the unplanned outage was both costly (e.g. $14,000 per minute x 55 minutes = $770,000) and potentially career limiting, to say the least.
In another situation at a large healthcare organization, a database record became misaligned and did not backup, resulting in a missed backup window completion and a missed SLA (service-level agreement). In this instance, the process fault not only impacted the enterprise, but the services provider too, resulting in severe penalties for both parties.
Keeping Data Safe
Verifying that data protection is meeting assigned service levels requires visibility and verification into whether all the data protection components operate as expected. The clarity gained by a single, unified view of the overall data protection environment health, performance, and reliability is necessary for proactively reporting protection to stakeholders and compliance to auditors.
Data Protection at Scale
Data protection management tools such as EMC Data Protection Advisor provide visibility and the insight necessary for data protection services whether deployed in enterprises or service providers supporting many enterprise or mid-sized customers. Data Protection Advisor 6.0 provides:
- Centralized management at cloud scale: Centralized management is not a new concept. The technology to host a centralized management function that can scale, however, is a more recent development. In Data Protection Advisor 6.0, load-balancing policies distribute information collected from supported devices across data stores on multiple hosts making it possible to scale the data protection service to support thousands of tenants on a single instance. It also enables greater granularity in reacting to changes. A new analysis engine correlates events from cross-domain devices with alerts in real-time making remedial actions possible before issues result in a missed SLA or incompliance—capabilities that could have saved money and careers for the financial institution and healthcare provider in the examples cited above. The financial institution would have gotten an alert about the outage in real-time, for example, with Data Protection Advisor.
- Multi-tenant capabilities: User dashboards are commonplace. The Data Protection Advisor user interface, however, adopts a horizontal, cross-domain approach to management more in keeping with cloud models, with an integration point for extending base capabilities through customization. A new user-interface based on the EMC Common User Experience (ECUE) design model provides a familiar dashboard look and feel for EMC customers using element managers like EMC Unisphere, as well as EMC ProSphere. The new dashboard provides individual views too for service provider customers giving tenants control over their own reporting, service-level monitoring, and show back—again, a base capability that could have kept the healthcare organization better informed. Multiple service rates can also be set and applied based on customer data protection needs or usage.
- Historical trending and analysis: Similar to previous releases, reporting capabilities include historical trending and analysis by tenant. In this release, however, a new REST API opens up many new possibilities by offering up the vast amount of real-time data collected by Data Protection Advisor to larger management frameworks for additional analysis. REST APIs are the new de facto standard for software technology deployed in cloud environments. Integrating Data Protection Advisor with a higher-level view of overall system health furthers service assurance across the entire data center. With this integration, the customers in my example
would have been able to examine recurring backup-related faults in relation to other activity to highlight high-potential problem areas, especially for infrastructure supporting mission-critical applications.
Cloud models coalesce around concepts such as compute at scale with the flexibility to extend capabilities and the ability to deliver IT-, platform-, or storage-as-a-service in real-time. This new paradigm requires management tools based on new architectures that can scale and work in these dynamic environments.
Data Protection Advisor 6.0 fulfills the needs of ensuring data replication and backup in multi-tenant cloud environments with a common user interface and an extensible REST-based API that fits with the other cross-domain management tools in the EMC portfolio. With the Software-Defined Data Center and Software-Defined Storage on the horizon, scalable management tools based on open architectures like Data Protection Advisor are probably the beginning of a trend.