Welcome!

Java Authors: Yakov Fain, Roger Strukhoff, Pat Romanski, Elizabeth White, Yeshim Deniz

Related Topics: Cloud Expo, SOA & WOA, Virtualization, Web 2.0, Big Data Journal, SDN Journal

Cloud Expo: Article

To Cloud or Not to Cloud?

Transformational metrics and analytics for managing "Infrastructure Anywhere"

Today's IT infrastructure is in the midst of a major transformation. In many ways, the data center is a victim of its own success. The growing number of technologies and applications residing in the data center has spawned increasing complexity, which makes IT as a whole less responsive and agile. While businesses are focused on moving faster than ever, large and complex infrastructure is inherently rigid and inefficient.

As a result, IT is moving outside the traditional data center into colocation facilities and cloud infrastructures - essentially Infrastructure Anywhere. The move to Infrastructure Anywhere is driven by the core objective of improving responsiveness and agility and reducing costs. For example, you can scale up resources through the cloud in minutes, not months. But for all of its benefits, this new Infrastructure Anywhere model presents critical challenges.

To make smart decisions about where to run applications and what kind of resources you need, you first must understand your workload: utilization, capacity, and cost. Gaining unified visibility is difficult when your application workloads are distributed across data centers and colocation facilities in different parts of the country or around the world. With limited visibility, how do you accurately align resources and capacity with workloads for efficient processing, cost control, and - more important - achieve the full business value of your IT investment?

It's All About Agility
According to the results of Sentilla's recent survey of data center professionals about their plans for cloud deployments, agility and flexibility are the top drivers behind enterprise IT transformation initiatives such as cloud deployments - followed closely by issues of capacity and cost.

Figure 1: Key drivers for cloud computing initiatives

While agility is the prime motivating factor, the importance of cost as a factor should not be ignored. According to the survey, the major resource limitation experienced by respondents - for all infrastructure initiatives - is budget.

Figure 2: Resource limitations

Note that several of the reported constraints (personnel, storage capacity) are related to the broader issue of budget. In this sense, cost is overwhelmingly the most important constraint on IT initiatives - including cloud initiatives.

2013 Is for Planning, 2014 for Deployment
Of the organizations surveyed, nearly 50 percent plan to be deploying cloud initiatives in 2014. Many are in the planning phases. In general, we can expect cloud computing deployments to increase by 70 percent in 12 months:

Figure 3: Data center cloud initiatives, by year

Similarly, those surveyed expect to gradually migrate more workloads to cloud platforms in the coming years - with 28 percent planning to run more than half of their applications in the cloud by 2014. The barriers to cloud migration are lowering.

Figure 4: Percentage of applications planned to move to the cloud, by year

The Cloud Isn't a Homogeneous Place
Cloud computing can refer to several different deployment models. At a high level, cloud infrastructure alternatives are defined by how they are shared among different organizations.

Figure 5: Where respondents planned to deploy cloud initiatives

Private clouds offer the flexibility of the elastic infrastructure shared between different applications, but are never shared with other organizations. Hosted on dedicated equipment either on-premises or in a colocation provider, a private cloud is the most secure but the least cost-effective cloud model.

Public cloud infrastructure offers similar elasticity and scalability and is shared with many organizations. This model is best suited for businesses requiring managing load spikes and scaling to a large number of users without a large capital investment. Amazon Web Services (AWS) is perhaps the most widely deployed example of public cloud infrastructure as a service.

Hybrid cloud offers the dual advantages of secure applications and data hosting on a private cloud and the cost benefits of keeping sharable applications and data on the public cloud. This model is often used for cloud bursting - the migration of workloads between public and private hosting to handle load spikes.

Community cloud is an emerging category in which different organizations with similar needs use a shared cloud computing environment. This new model is taking hold in environments with common regulatory requirements, including healthcare, financial services and government.

The research showed that organizations are evaluating a broad range of different cloud solutions, including Amazon AWS, Microsoft Azure, Google Cloud Platform, and Red Hat Cloud Computing, as well as many solutions based on OpenStack, the open source cloud computing software.

Without planning, ad hoc cloud deployments combined with islands of virtualization will only add complexity to the existing data center infrastructure. The resulting environment is one of physical, virtual and cloud silos with fragmented visibility. While individual tools may deliver insight into specific parts of the infrastructure puzzle (physical infrastructure, server virtualization with VMware, specific infrastructure in a specific cloud provider), IT organizations have little visibility into the total picture. This lack of visibility can impede the IT organization's ability to align infrastructure investments with business needs and cost constraints.

Infrastructure Complexity Is the New Normal
While it aims to bring agility to IT, the process of cloud transformation will only increase infrastructure complexity in the near term. IT organizations must manage a combination of legacy systems with islands of virtualization and cloud technologies.

When asked about where cloud infrastructure will reside, survey respondents indicated that they will be managing a blend of on-premises and outsourced infrastructure, with the balance shifting dramatically from 2013-2014.

Figure 6: Where cloud infrastructure will reside, by year

The Need for Unified Visibility into Complex Infrastructure
As you plan your own cloud initiatives, you must prepare for multiple phases of transformation:

  • Deploying new applications to the cloud as part of the broader application portfolio
  • Migrating existing applications to cloud infrastructure where possible and appropriate
  • Managing the hybrid "infrastructure anywhere' environment during the transition and beyond

To support these phases, you need visibility into workloads and capacity across essential application infrastructure - no matter where it resides. From the physical and virtual resources, up through applications and services, you will need insight so you can align IT with business objectives.

Figure 7: The need for infrastructure insight at all levels

Essential Infrastructure Metrics for Right-Sizing Infrastructure
Decisions about which applications to deploy to the cloud and where to deploy them will require visibility into:

  • Historical, current and predicted application workload
  • Current and predicted capacity requirements of the workload
  • Comparative cost of providing that capacity and infrastructure on different platforms

For application migration scenarios, you will need to understand the actual resource consumption of the existing application. Whether it's a new application or a migrated one, you will need to ‘right-size' the cloud infrastructure to avoid the twin dangers of over-provisioning (and wasting financial resources) or under-provisioning and risking outages or performance slow-downs. You will need insight into:

  • Memory utilization
  • CPU utilization
  • Data transfer/bandwidth
  • Storage requirements

You will also need good metrics about the cost of running the application in your existing data center, as well as the predicted costs of running that same application on various platforms. These metrics need to factor in the total cost of the application, such as:

  • Personnel for supporting the application
  • Operating system
  • Management software
  • Cooling and power
  • Leased cloud
  • Server and storage hardware

To accurately predict the cost of running the application on cloud-based infrastructure, you will need accurate metrics around the actual, historical resource consumption of the application (storage, memory, CPU, etc.) as it maps to billable units by the provider. By understanding the actual consumption, you can avoid over-provisioning and overpaying for resources from external providers.

Infrastructure Analytics for Cloud Migration
For any given application that is a candidate for the cloud, you want to be able to compare the total cost of the resources required across different options (public, private and on-premise).

While you could try to manually crunch these numbers in a spreadsheet, the computations are not trivial. These are decisions that you will need to make repeatedly, for each application, and revisit when an infrastructure provider changes its cost model or fee structure. For that reason you'll want a tool that lets you get an accurate and continuous view into current costs and model "what-if" scenarios for different deployments.

Continuous Analysis for Continuous Improvement of the "Infrastructure Anywhere" Environment
Before deploying an application, what-if scenarios help you make sound resource decisions and right-size applications. After deploying, continuous analysis is key to ensuring that you are optimizing capacity and using resources most efficiently.

While individual tools may already give you slices of the necessary information, you need integrated insight into the complete infrastructure environment. Again, emerging infrastructure intelligence can assemble necessary information from applications and assets that are both on and off your premises, virtualized and not, in different platforms and locations.

Figure 8: Transformational analytics

The software can provide 'single pane of glass' visibility into assets and applications throughout the physical, virtual, and cloud infrastructure, including:

  • Application cost/utilization spanning different locations
  • True resource requirements of apps (for more accurate provisioning in cloud infrastructure)
  • CPU and memory utilization of apps, wherever they reside

Summary
By 2014, enterprise computing will look quite different than it does today, yet many legacy systems and infrastructure will still be with us. IT operations, business units and application architects will need to manage applications that reside in infrastructure that spans on-premise and offsite locations, with public, private, hybrid, and community cloud infrastructure. Data centers will be just one part of the total pool of infrastructure that IT manages on behalf of the business.

To manage this transformation, you will need to make smart decisions about where workloads should reside based on specific application and business needs. As these changes roll out, you will need to manage the transforming and hybrid application infrastructure to deliver the necessary performance and service levels, no matter where applications reside.

IT organizations need the insight to make fast, smart and informed decisions about where workloads and data should reside and how to deploy new applications. Rather than isolated silos of metrics and capacity and utilization data, IT needs unified visibility into infrastructure across the virtual computing environment - both on-premise and off. And they need the metrics and continuous analysis to manage the evolving infrastructure in a manner aligned with business objectives.

An emerging category of infrastructure intelligence can provide the continuous and unified analytics necessary to understand and compare your decisions and to manage the data center during the transformation. With broad 'infrastructure insight' you can align cloud platforms with business needs and cost requirements - delivering the agility to realize new revenue opportunities with the insight to contain the costs of existing applications.

More Stories By Ranvir Wadera

Ranvir Wadera is senior vice president of product development at Sentilla Corporation, an Infrastructure Intelligence software platform provider based in Redwood City, California. He has more than 20 years of product development and management experience in both entrepreneurial and large company environments. Before joining Sentilla, he was founder and CEO of Rjenda, a SaaS solution for the education market. Prior to Rjenda, he was vice president of product development for query and reporting tools at Hyperion/Oracle. Prior to Hyperion, Ranvir was vice president of product development at Business Objects. Early in his career, he also held product development positions at Oracle and developer and manager roles in India.

Ranvir received his undergraduate and post-graduate education in India from the University of Delhi and the International Management Institute, with follow up technical education from University of California at Berkeley.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.