Reaching to the Cloud

Business intelligence (BI) and cloud computing are starting to look like a match made in heaven.

Firms of all sizes are recognizing the power of BI to find markets, identify savings, and make the most of opportunities. But BI solutions put heavy demands on IT infrastructure. On top of licensing costs, expanded infrastructure is a heavy burden for firms to bear up-front.

The good news is that cloud computing, especially software as a service (SaaS) is coming to the aid of BI, making "BI for rent" affordable for firms of all sizes. BI from the cloud means that the user doesn't have to buy or maintain a complex infrastructure. The vendor or cloud provider takes care of all that. Users only pay for what they need, when they need it.

Which is why BI from the cloud has become the Next Big Thing. Cloud BI offers three huge advantages over licensed and locally-hosted approaches:

  • Scalability – Cloud-based software-as-a-service (SaaS) solutions can easily be adjusted to meet the requirements of companies of all shapes and sizes. Even the smallest businesses can tailor a cloud BI system to fit their needs and budget..
  • Ease of Entry – Traditional site-licensed software imposed heavy upfront IT costs to get the system up and running. But with a cloud-based BI solution, that upfront cost is reduced because all hardware and software systems are located and managed remotely by vendors and service providers.
  • Cost – Cloud-based systems allow companies to "rent" the software as opposed to buying a full license. This subscription-based model ensures that firms only pay for their organization’s actual usage.

 

Moving to the Cloud

Adopting cloud BI does have its challenges. Whether a firm's database facilities are large data centers or a few servers in a back room, the data will be need to be migrated. Most firms will find that this is a good time to call in professional movers.

Successful data migration involves several steps:

Assessment

The first step in migrating is assessing specific needs. This means taking a full inventory of the application portfolio, to determine the impact of database migration on the IT ecosystem. This assessment will consider other applications, integration, reporting, and data backup and recovery.

  • Factors driving the migration.
  • An inventory of the current environment, including programs, scripts, and external interfaces. Also included is information on hardware and software configurations, including operating system and database versions in use.
  • Migration tools and implementation options. A carefully designed and highly automated migration process can save time and money spent on migration and testing.
  • Determination of IT resources needed for the target database, and resources needed to support the migration project.

Analysis and Design

The implementation details of the target database must be determined and specified. This covers not only data types and SQL code, but also transaction management and security roles and privileges. The migration plan must support needed functionalities and features. All this involves:

  • Determining the database schema layout. Databases vary in how they organize and handle schemas, objects, and users. Differences in schemas between source and target databases can produce object collisions, since different databases allow objects with the same names but different data structures.
  • Mapping data types. While basic data types for numerical, character, and timestamp data are fairly standardized, significant differences can exist in data length limits and resulting data precision.
  • Database locking behavior. Databases under heavy use loads may display lock escalation behavior.
  • Important little details. Case sensitivity or insensitivity, and use of zero-length strings can lead to serious complications if not taken into account.

The Actual Migration

Migrating from one relational database to another is (relatively!) simple, because the objects in a relational database are organized in fairly standard ways. Things get more complicated for migrations of non-relational databases, such as network or hierarchical databases, whose structures are more varied.

  • Migration of database schema. The migration process must ensure that all objects are transferred from the source to the target database, along with users' roles and privileges with respect to objects. Table portioning options must be considered.
  • Data migration. When migrating very large databases – those with several terabytes of data – it is essential to have a sound data migration strategy in place. This means having appropriate tools available, and in particular making effective use of database features such as compression and partitioning.
  • Database stored program migration. Migration will include stored procedures, triggers, and views which, in many relational databases, are used for implementing critical business logic. Steps may include cleaning up and optimizing code, testing error handling, and reviewing code in Java or other languages.
  • Application migration.  The biggest migration challenge is moving customized applications, or those developed in-house, often built around the source database.

Migration Testing

The migration must be tested as it is executed. This testing usually involves not only data verification, but also testing of migrated business logic embodied in stored functions, procedures, and triggers. Application interactions with the new database must be tested, along with database maintenance scripts.

  • Data verification must closely monitor the migration process and ensure that it is error-free.
  • Stored functions and procedures must be tested once migrated.
  • Applications must be tested on the new database.

Post-Migration Optimization

Applications that have been migrated from one database to another one may show degraded performance. This typically happens because the application was designed around the initial database, or has been tweaked over time to optimize it for that database. Application performance must therefore be tested, and applications re-optimized to minimize bottlenecks. Factors to look for include:

  • Insufficient system resources provided by the cloud.
  • Process of query executions.
  • Larger workloads, possibly due to concurrent user populations.
  • Database structures lack sufficient capacity.

BI Cloud Options

BI from the cloud is not just offered by multiple vendors – it comes in a variety of flavors to meet the needs of individual firms. Some of these are:

  • Cloud Infrastructure, or infrastructure as a service (IaaS). In this option, the user "owns" and fully controls the BI solution package, but "rents" the infrastructure environment that supports it. This option provides the advantages (but also the responsibilities) of full solution management, alongside the savings of rented infrastructure environment.
  • Cloud platform as a service (PaaS). In this option the cloud vendor provides a broader range of support, while the user retains control of the BI core application. This option offloads more support to the cloud vendor, while retaining significant management of the BI solution.
  • Cloud analytics, or software as a service (SaaS). In this option the user rents a complete BI service package. The solution comes off the shelf, with limited user management options, but a maximum of vendor support.

These cloud options really represent a continuum of cloud BI approaches, allowing firms to select the mix that best fits their specific needs.

Reasons to Adopt Cloud BI

The reasons for reaching to the cloud for BI all come down to one: ROI. But every firm's needs and strategies are different, and companies are finding a variety of approaches to cloud BI. These include:

Coordinating cloud resources with on-site resources. For companies with a major IT presence, a mixture of cloud and local applications and capabilities may be optimal.

Internal consolidation. Some large firms may be pulling together and coordinating BI projects that originated separately in various departments. A private cloud can allow these separate sub-units to share licensing costs and maximize utilization of resources.

Fully in the cloud. For many other companies, cloud BI goes hand in hand with a full suite of other cloud-based services such as customer relationships management (CRM). A variety of business needs can favor a fully cloud-based BI strategy.

  • SMBs. Small and midsize businesses want access to the power of BI, but cannot afford to implement complex BI technology in-house. These firms are candidates for fully cloud-based BI.
  • Rapid Growth. Fast-growing firms need room to grow, and might quickly overtax an in-house solution. Again, the cloud provides ample room to grow.
  • Elastic Demand. Many firms have highly seasonal or other varying capacity needs. These firms don't want to invest in resources that sit idle much of the year, making renting from the cloud an attractive alternative.

Challenge and Opportunities

Moving to the cloud is not without its challenges. For many firms, moving critical data and applications out of the local environment raises important security concerns. Moreover, cloud technology is still new, and for many firms poses the challenges of new ground. And, beyond the initial migration concerns outlined above, companies need to consider the speed and reliability of connections among multiple cloud vendors and resources.

  • Security – For many businesses, questions and concern still remain about the physical, network and external security of cloud computing platforms.
  • Emerging Technology – Cloud computing is still evolving, and companies must expect and adjust to unforeseen and unexpected glitches as the web-based platform matures.
  • Data Movement – Companies will have to rely on data movement outside of their "four walls." In contrast to an in-house system where connection time and access is typically reliable, cloud-based products depend on multiple systems (i.e. a vendor’s servers, ISP uptime, etc.) to complete day-to-day BI tasks.

In spite of these challenges, the flexibility, freedom to innovate, and above all the cost advantages of cloud BI add up to a compelling argument for the leap to the cloud.

The leap is much easier if it is guided by specialists with a background in BI, security, infrastructure, and the other components of a cloud migration. GRT Corporation brings 17 years of technology experience, and a record of showing firms large and small, across industries, how to meet their specific IT needs.

At GRT Corporation our focus is not on selling particular vendors' products, but entirely on our customers, and their technology requirements and objectives. Let us be your partner in reaching for the power of BI in the cloud. 

 

To learn more about how GRT can help you meet your goals, contact us.