In this second article on the excellent Cloud Connect 2013 conference, we look at selected sessions that should be of keen interest to IT managers and enterprise cloud customers. Key findings are presented in a concise form to aid readability. We then report the results of the Enterprise Cloud Adoption survey by Everest Group.
The third article in this series will summarize leading-edge Cloud Research and Analysis by McKinsey & Company and also look at disruptive aspects of cloud computing, as discussed in several Cloud Connect 2013 panel sessions.
Effective Cloud Network Deployments, Monitoring & Optimization:
In a presentation, “Effective Networking Strategies for Your Cloud Deployment,” Eric Hanselman of the 451 Research Group suggested that cloud users should first establish networking requirements in terms of the metrics important to their organization. These may include: performance requirements, data flows, data paths, application segmentation, user-to- application paths and direction, intra-application volume and performance.
Network options involve varying levels of sophistication, including L2 (Data Center switching and Carrier Ethernet bridging ) and L3 (IP VPN routing). External connections should be based on requirements for aggregate bandwidth (throughput), latency, and L2/L3 path control. access control lists, firewall placement, encryption, VPN management, prioritization, and optimization (compression, caching and reformatting) all need to be considered as well.
Amazon’s Virtual Private Cloud (VPC) offers more structure, according to Eric. It includes: network segmentation, routing tables (with multiple interfaces per instance), virtual appliance support (with firewalls, intrusion detection systems and APM/WAN optimizers). Rackspace and Google cloud networking offerings were also profiled (but not Savvis/Century Link which offers many networking options including a customizable IP-MPLS VPN).
Sharon Wagner of Israel- based Clodyn provided lessons learned from AWS cloud deployments in his talk titled, “Best Practices in Cloud Optimization.” Some key findings:
- 58% of IT shops run unmanaged cloud applications
- Cloud applications change every second week
- Nearly 13% of the configuration is changed in every release (of cloud application software)
- Around 80% of the instances are significantly (~12%) under-utilized
- Customer reports indicate that 15% over-utilized cloud apps causes a 7% revenue loss
Such dynamic cloud environments results in over provisioning, wasted resources and budget violations. The solution is exactly what Cloudyn does- it monitors, analyzes, diagnoses, and optimizes cloud deployments. Effective cloud deployments were said to require consistent monitoring and optimization of usage, performance, and cost.
Life-cycle, and usefulness to the business should also be evaluated, according to Mr. Wagner. The only metric that matters is the impact to the business, i.e. Return on Capital Investment and that can only be achieved by making cloud operations more efficient.
Enterprise Cloud Adoption Survey:
Cloud Connect and Everest Group conducted a joint survey in 2012 and 2013 to identify cloud market trends and disconnects. The objectives of the surveys were to:
- Identify broad based cloud adoption patterns
- Identify barriers to adoption
- Evaluate overall “cloud sentiment” post adoption
“Year over year, we are seeing the emergence of next generation budgets within business units,” said Scott Bils, partner at Everest Group. “Cloud buying power is concentrating in lines of business, not the CIO’s office.”
Business buyers are the primary cloud adoption drivers. It’s the business stakeholders who are the primary decision-makers for most major enterprise workloads. They’re primarily concerned about speed, flexibility, scale and reduced time to market. Those are the top cloud adoption drivers. Surprisingly, cost savings was not a significant adoption factor.
Here’s a summary of the key findings of the first survey, which was completed in October, 2012:
- Cloud adoption is expanding beyond “low hanging fruit” of SaaS, such as email and custom applications.
- Buying focus is turning away from horizontal use cases towards more customized business applications.
- Cloud is seen as an enabler of topline growth, beyond the cost reduction imperative.
- Perceived “security issues” continues to be the most significant barrier to cloud adoption.
- While VMware has leading mindshare (for IaaS), many enterprises prefer open source cloud software (e.g. OpenStack).
- Cloud Service providers need to adapt to new budget centers and sell on business value rather than cost.
- SaaS modules have been adopted most widely; IaaS adoption is expected to grow fastest in the near future.
- Overall, buyers’ “cloud sentiment” remains extremely positive, with high expectations for the future.
The focus of the most recent survey (first half of 2013) was on private vs public clouds. Not surprisingly, private cloud is still the overwhelmingly preferred choice by survey respondents! Enterprises still show a preference for private cloud models for most workload types. “This increasing preference towards private cloud models within enterprises could potentially be driven by internal ‘IT marketing,’” said Bils.
Note: This author has long believed that security issues and variable performance (depending on aggregate workloads and network loading) have hindered adoption of public clouds for quite some time. That’s despite all the press about Amazon AWS, Rackspace, Google, Microsoft, AT&T and other public cloud providers. The Metro Ethernet Forum (MEF) is focusing on Carrier Ethernet for delivery of Private Cloud Services, rather than public clouds.
The tipping point for enterprise clouds may be here. Everest Group found that the majority of enterprises now expect migration to some type of cloud delivery model across all major workload types. This is illustrated in the adjacent figure.
The “Private Cloud Infrastructure Wars” are heating up, according to the Everest Group survey. While VMware is the current leader, a significant proportion of the market professes to be patform agnostic or prefer open source platforms, such as CloudStack and OpenStack. Please see the figure below for more details. Note that Public Cloud Infrastructure is solely determined by the CSP, e.g. Amazon AWS.
Growth in adoption of cloud services is creating two distinct, but overlapping buying groups in the enterprise as shown in the chart below.
Image Courtesy of Everest Group
Further Information from Everest Group:
Readers may download a copy of the 2013 Enterprise Cloud Adoption Survey Summary Report at: http://everestgrp.com/ccevent
7 Things We Learned at Cloud Connect | Gaining Altitude in the Cloud, by Scott Bils of the Everest Group is at:
Stay tuned for 2013 Cloud Connect Part III! Here’s a preview: McKinsey consultants Will Forrest and Kara Sprague proposed a different, and enormously disruptive, scenario of the ultimate cloud adoption road map. Like IDC, McKinsey sees the future of IT in public cloud computing, which may very likely be in the form of a separate IT organization, created specifically to reside outside of the existing one. Randy Bias of Cloud Scaling and other panelists detailed a laundry list of disruptions brought on by cloud computing which we will share with Viodi View readers.