When Clouds Get Cloudy
Growing complexities in cloud computing
Oct. 26, 2013 04:00 PM
The worldwide market for cloud computing is rapidly growing across all segments - 36% CAGR through 2016, reaching a market size of $19.5B by 2016, according to the InfoPro Wave 5 Cloud Computing Study from 451 Research. The result has been a substantial increase in the operational complexity of IT infrastructure management.
Legacy IT service management (ITSM) tools and suites from vendors like IBM, HP, BMC, CA and Microsoft, designed for on-premise data centers, are no longer effective for managing cloud-centric applications and infrastructure. Therefore companies are adopting multiple open source and point solutions designed to manage specific aspects of their cloud infrastructure.
The narrow focus these new products and services deliver enables companies to assemble best-of-breed "ITSM-like" solutions, but their lack of integration has resulted in fragmentation of critical operational and business data across their environment. The challenge now becomes - how to take these complex data sets from disparate tools and services and obtain an integrated view that enables critical business decision-making around topics such as cost, availability, performance, and security?
ITSM - Changing the Landscape with Cloud
We are in the early stages of a major technology shift that is disrupting the traditional ITSM suites. Early cloud adopters have quickly learned that the traditional suites were ineffective for managing highly dynamic and complex cloud infrastructure. Instead they opted to assemble a collection of more modern and less integrated point solutions with names like Chef, Puppet, Nagios, Sensu, New Relic, Splunk, PagerDuty, Pingdom, and Airbrake. These new products and services began to take hold in IT operations, and have today replaced the previous generation of ITSM solutions within most organizations.
There are four major factors driving the abandonment of the legacy ITSM suites:
- Suite stagnation: Many of the products that manage enterprise data centers were developed in the early and mid-1990s. While vendors have continued to invest in these products, their age and bulk has resulted in a steady slow down in innovation.
- Point tool innovation: New point tools such as Splunk and New Relic were designed in the post-SaaS era, and therefore have provided a fresh perspective on traditional ITSM challenges. These products are much easier for customers to evaluate, purchase, deploy and operate than the legacy ITSM suites.
- Cloud complexity: Cloud computing has introduced several new innovations that have stretched the capability of the ITSM suites, including on-demand infrastructure, consumption-based pricing, platform services, infrastructure as code, and elastic infrastructure. The legacy ITSM suites were designed for a world of physical hardware, capital expenditures, and less frequent change.
- Price: The legacy ITSM suites were priced for enterprises with few alternatives. The new point tools are often are delivered via SaaS, priced monthly, easy to try, and in some cases even distributed as open source.
The Impact of Growing Complexity in the Cloud
While these new point products have enabled early cloud adopters to successfully build and manage large-scale infrastructures, a new challenge has surfaced: pulling the disparate and disconnected portals into a consolidated view required to run a business. There are three primary challenges that organizations face in consolidating their virtual ITSM suites: complexity of data integration, the transient nature of cloud infrastructure, and complexity of pricing.
- Complexity of ITSM data integration: While the new point solutions enable incredible power and control, they are typically narrowly focused and lack a cohesive system view. IT, CFO, CEO, and other stakeholders often feel acute pain from the lack of cohesive view across the organization. Most cloud customers eventually realize the need for an easy, repeatable, auditable method to holistically view their infrastructure to make critical operations, engineering or financial decisions.
- Transient Nature of Cloud Infrastructure: The cloud promotes the use of disposable infrastructure. It is very common to run large amounts of infrastructure for short periods of time, or to frequently launch new infrastructure to replace existing infrastructure. While this is highly efficient, it results in short-lived infrastructure often in support of long-lived business needs, increasing the challenge of pulling together an integrated view for decision-making.
- Complexity of Pricing: To further complicate matters, public cloud vendors offer multifactorial pricing plans that are often designed to socially engineer specific usage patterns of their services. Understanding and managing to the most efficient usage of these plans is essential for cost efficient operations of public cloud infrastructure.
As the shift to cloud computing continues to gain momentum, individuals responsible for cloud infrastructure are faced with a near impossible task - how to embrace the advantages of the cloud without giving up the ability to make informed business decisions?
Why Integration Matters
In order to gain business insights across people, processes and technology, organizations needs a holistic view of their applications, infrastructure, and business. They need to be able to answer questions such as:
- How do you report on what is happening inside the infrastructure?
- How do you identify usage of applications?
- How do you identify and optimize costs?
- How do you optimize workloads?
- How do you view and optimize cloud usage by customer, service, product, or department?
- How do you consolidate your fault, configuration, accounting, performance, and security data of your cloud ecosystem?
Challenges of Custom Solutions
The first wave of adopters who have embraced cloud computing have typically been staffed by forward thinking technical teams that are capable of stitching together minimum solutions for integrating disparate tools within their organizations. But as the cloud achieves early majority adoption, organizations have found less time, capability and tolerance to make the investment required for custom solutions. The result is that the driving benefits of the cloud - consumption-based pricing, on-demand infrastructure, and business agility - often become negatives as organizations scale their usage.
In the early phases of growth, organizations often rely heavily on manual labor to pull together data from multiple sources. Spreadsheets are typically the primary integration tool used to slice and dice information in support of decision-making. While this can provide a passable solution at moderate scale, it is often highly resource intensive, error prone, and time consuming.
As enterprises continue to grow their cloud usage, they often reach a tipping point at which they feel acute pain from the lack of integration across numerous point tools. Often this results in poor or inefficient usage of the cloud, forcing organizations to make critical business decisions with insufficient data. Some organizations become cloud dropouts at this stage, returning to the safety of physical infrastructure. A few technical pioneers (e.g., Netflix, Heroku), have found themselves creating expensive and customized internal systems and open source to tame the cloud.
Unfortunately most enterprises do not have the time or expertise to develop custom internal solutions to manage their cloud infrastructure.
The Next Generation of ITSM
How do you embrace the advantages of the cloud without giving up the ability to make informed business decisions?
The ability for organizations to successfully harness the cloud will hinge upon an emerging category of new cloud analytics platforms that enable a best of breed solution to ITSM. Unlike their predecessors, these new ITSM platforms will need to be cloud centric (capable of handling volatile, diverse cloud data sets) and API driven to facilitate integration with best in class open source products.
At a minimum the next generation of ITSM must deliver:
- Consolidated data aggregation and correlation
- Detailed asset management across all core system management products - assets, log management tools, automation and provisioning tools
- Usage, performance, and financial metrics across the cloud ecosystem
- Long term storage for the data to deliver trended reporting
- Audit trails for change management and security
- Service level reporting by functional business group
Until we successfully deliver a holistic view of the cloud ecosystem, organizations will remain challenged by the growing complexities of scale in cloud environments.