Success by Design in cloud data management

InsightaaS: As businesses move past the exploratory stages of cloud into its use as a production-grade component of corporate infrastructure, we are seeing new cloud management issues emerge. Most organizations have moved beyond "should we use cloud?" and "should we look at IaaS, PaaS and/or SaaS" to issues associated with enterprise-level IT operations. Some of these are related to new compute management issues: "how do I get the best performance from my on- and off-premise resources, and how do I impose consistent management, security, and back-up practices across these environments?" Other questions relate to the ‘plumbing’ associated with cloud: "how do we quickly and safely move data across environments?" Still other questions arise from the options that cloud presents — for example, "in a big data world, how do we evolve from seeing data as an application output to seeing data as an asset, and what does that mean to our management strategy?" and, "how do we connect our cloud-based capabilities to support new workloads or work modes, such as analytics and/or mobility?"

Class 395 Bullet train
A Class 395 bullet train. What's more impressive: 225 km/h or 1 petabyte/day?

Paul Lewis brings an informed and engaging perspective to these issues. As the former vice president, technology for Davis + Henderson, he understands IT challenges from a buy-side perspective, and as current CTO for Hitachi Data Systems (HDS), he has deep insight into the technology options available to IT executives.  In the wake of HDS’s introduction of the Hitachi Content Platform (HCP), a converged content platform offering unique web data management capabilities, we renewed our discussion with Lewis to look at how HCP — and converged systems generally — help IT management to address issues associated with production-level cloud use.

We have published an edited transcript from our discussion (it's available here), to provide a detailed review of Lewis's thoughts. To provide a sense of its content, we've extracted six highlights from the interview:

  1. Keys to enabling data for use by mobile systems: Lewis views this as a two-part problem. The first issue is abstracting the data away from the application (e.g., PowerPoint, a mortgage application) and underlying platform infrastructure, and the second is making the data available for use by mobile users — making it possible for mobile applications to combine data from multiple sources in ways that make it useful to the mobile user.
  2. Metadata is the key to managing combinations of structured and unstructured data: Lewis states that "any object platform (such as HCP) that creates metadata that surrounds files" provides the tools needed to correlate diverse files and file types. "The complexity," he says, comes largely from applications that lack an integrated feature supporting independent use of data, requiring IT management to use systems like HCP to abstract information and wrap it in metadata that enables use of the data in other contexts. Once the metadata is in place, IT management can apply business rules governing issues like data access, security, archiving, and the creation of audit trails.
  3. Tiering storage into the cloud, and use of cloud-based "data lakes": Lewis emphasized that organizations that deploy a single tier of storage are using "the most expensive" approach to storage. He urges adoption of a "data life cycle" governing storage from the time that a file is created to the time when it should be removed. Lewis quotes research showing that "40 percent of data isn't looked at after 3 months and 70 percent of data, especially unstructured data, isn't looked at after one year." These aging files should be moved to lower-cost storage, and/or to cloud-based "data lakes" where files that need to be kept to meet regulatory requirements can be stored at very low cost.
  4. Priority executive issues include cost, flexibility, and security: In the interview, Lewis described the issues that are central to his storage-related discussions with other IT executives. Managing total cost (through tiering) is important, as is providing alternatives to Dropbox or Google Docs that allow users to synchronize data while still retaining corporate control of that information: "controlling [data] within your domain, within your security profiles, within your active directories, and with your policies, is critical from a security perspective." Flexibility can be interpreted in two ways — flexibility in terms of use and deployment (e.g., software-only vs. existing infrastructure, capture of data into a central repository that is not dependent on specific applications or architecture), or in terms of financial models (CAPEX vs. OPEX, leasing).
  5. Converged infrastructure supports many different use cases — which improves investment payback: Lewis noted that he has seen HDS customers use HCP for archiving ("where you don't want to apply complex archiving policy and implementation to your 400 or 500 applications"), for file sync/share (through the HCP Anywhere functionality), and as a means of supporting a Big Data repository (including the attachment of metadata) — and as Lewis points out, "most of our clients use it for more than a single purpose, so it's not just for file sync and share, it's used for the grander purpose of archive, file sync and share, analytics, data mobility, or even just simply making structured what inherently is unstructured." Converged infrastructure allows customers to gain better leverage on investments by enabling them to gain incremental functionality from foundational technology. Lewis went on to note that customers purchase HCP to address a single point of pain, but that "they wouldn’t buy it for a point solution, they would buy it knowing that their road map includes implementation of more than one workload, delivering more than one source of value."
  6. Enabling CIO evolution: In the final section of the "Success by Design in Cloud Data Management" transcript, we capture Lewis’s thoughts on how CIOs can evolve from a ‘keep the lights on’ mentality to a focus on creating new value. Lewis believes that this transition "requires a different appreciation for the assets that you have, a different appreciation for how you're organized and a different appreciation for where you should be allocating time and attention." HDS works with CIO clients to create strategies that will carry them through this evolution, focusing on the technology steps required to create "unexpected new value, above and beyond just simply keeping the lights on."

Read the entire interview transcript (no registration required): Link