Planning cloud analytics? Have you considered…

A holistic approach for planning analytics in the cloud is facilitated by asking relevant, comprehensive questions as well as being familiar with components that are essential for implementing cloud analytics.

As explained by Lakshmi Randall, Research Analyst, Gartner, these essential components (analytics platform components plus supporting components) are listed below:

Analytics Platform Components:

• Business Intelligence and Analytics, data discovery, advanced analytics, industry-oriented prepackaged analytical applications

Supporting Components (not intended to be an exhaustive list):

• Data Integration supporting analytics solutions

• Analytical data stores (data warehouse, data mart, data lake etc.)

• Data Preparation supporting analytics solution

• Data Quality supporting analytics solutions

• Identity and Access Management

Key questions to ask when planning a Cloud analytics solution are offered below; this list is not intended to be exhaustive, explains Randall.

These planning questions are predicated on placing the analytics platform components in the Cloud; the locations of supporting components are then determined by the strategists, planners and architects based on considerations such as location of data sources, an enterprise’s investment in existing on-premises architecture, and budget and staffing constraints.

“When used in conjunction with business objectives - modernisation, reduction in cost or time to market, reduction of on-premises infrastructure footprint, support of innovative business use cases - Randall believes these questions can serve as a starting point for developing a holistic approach to Cloud analytics.

“Note that it is imperative to consider the cost and staffing (skillsets, maintenance, services) ramifications of the responses to each question,” Randall adds.

Requirements planning:

1. Have you identified your business requirements (e.g., analytical model, KPIs, measures, reports, and dashboard requirements)?

2. Have you identified the end-users (publishers and consumers) of the cloud analytics solution?

3. What analytic capabilities are you planning to implement (such as Data discovery, Business Intelligence and Analytics, Industry-oriented prepackaged analytical applications, Advanced Analytics)?

4. Is your cloud analytics solution internal facing or external facing?

5. Are you planning to leverage relational, non-relational (NoSQL, Hadoop), or both data stores in the cloud platform in support of the analytics solution?

6. Have you identified governance, quality and integration requirements?

Page Break

Governance planning:

1. What are your requirements for data security and data privacy?

2. Is the identity management and access solution for Cloud analytics federated with the on-premises solution, or is it a standalone solution in the Cloud?

3. Is it necessary to perform data anonymization and masking before, or while moving the data to the cloud platform, or when exposing the data for cloud access?

Location-specific planning:

1. Where do/should the analytical data stores reside? On-premises, Cloud, Hosted, or combination thereof?

2. Where do the data sources leveraged by the data integration process (to populate the analytical data stores) reside? On-premises, Cloud, Hosted, or combination thereof?

3. If there are data sources directly consumed by the analytics solution (thus bypassing analytics data store), where do these reside? On-premises, Cloud, Hosted, or combination thereof?

4. Where do/should the data integration services reside? On-premises, Cloud, or Hybrid (Cloud combined with on-premises)?

5. Where do/should the data quality services reside? On-premises, Cloud, or Hybrid (Cloud combined with on-premises)?

6. Where do/should the data preparation services reside? On-premises, Cloud, or Hybrid (Cloud combined with on-premises)?

7. Where do/should the HA/DR solutions reside? On-premises, Cloud, Hosted, or combination thereof?

Cloud Services planning:

• For analytics platform and supporting components (such as data integration, identity and access management, analytical data stores) residing in the Cloud,

• Are you considering Public Cloud, Private Cloud, or Hybrid Cloud (i.e., either Public or Private combined with on-premises)?

• Which Cloud service implementation model are you considering; Platform as a service (PaaS), Infrastructure as a service (IaaS), Software as a service (SaaS)?

• Are you considering a single CSP (Cloud Service Provider) or multiple CSPs for implementing your Cloud analytics solution (e.g., one Cloud provider for the analytics platform and a different Cloud provider for data integration)?

• How are you planning to extend the corporate data center perimeter to the Cloud (see Hybrid Cloud Network Architecture)? Have you consulted with the security organization on this effort?

• What are your requirements for scalability, elasticity, and performance?

• What are your Service Level Agreements (SLAs) for the Cloud analytics solution?

By Lakshmi Randall - Research Analyst, Gartner