Yi9x6RMc

EMA Research Report: Optimizing the Network

for Reliable Application Delivery Across the Cloud

Executive Summary

EMA research has documented strong and growing interest in cloud deployment models. While private

cloud received the earliest attention, stronger adoption is now happening with public and hybrid models.

In EMA’s April 2014 report, “Managing Networks in the Age of Cloud, SDN, and Big Data: Network

Management Megatrends 2014,” over 50% of respondents reported public/hybrid cloud initiatives

were driving network management priorities. Comparing the results from the Megatrends 2012 study

to 2014, cloud projects have moved from early-adopter status to mainstream business initiatives, and

their impact on network management has grown from 36% in 2012 to over 50% in 2014. However,

cloud deployments are not without their challenges. The EMA Network Management Megatrends

2014 study revealed that 75% of the companies that had deployed in the cloud reported difficulty

with monitoring and troubleshooting their cloud deployment. Cloud represents a new infrastructure

paradigm, rendering traditional monitoring, control, and optimization methods less effective, or even

completely ineffective. EMA seeks to better understand what steps are being taking to improve visibility

and performance of cloud deployments. This study looks at what types of workloads are being placed

in the various cloud deployment models, examines how performance issues are being dealt with, and

considers whether cloud benefits are being fully realized.

The Shift to Cloud

Moving workloads to the cloud has the potential to reduce operational overhead and reduce CapEx

spending, but only if the benefits of the cloud are fully realized. Cloud computing comes with its own

unique set of challenges, not least of which is the splitting of workloads between internal and external

resources. The purpose of this report is to better understand how performance, and to some degree

monitoring, of cloud deployments is being handled and to see how both performance and monitoring

are affected by cloud type.

Since “cloud computing” can mean different things to different people, for the purpose of this survey

EMA used the following definitions:

• Private Cloud – The private cloud infrastructure operates solely for a single organization and is

managed and hosted internally, inside the corporate firewall.

• Public Cloud – The public cloud infrastructure is provisioned for shared use. It may be owned,

managed, and operated by a business, an academic or government organization, or some

combination of those organizations. The cloud service exists on the premises of the cloud provider,

outside the corporate firewall. Public cloud includes services such as Amazon Web Services (AWS),

Microsoft Azure, IBM/SoftLayer, and Google Compute Engine.

• Hybrid Cloud – The hybrid cloud infrastructure is a combination of public cloud services and an

on-premises private cloud. These systems are tied together via APIs or other proprietary technology

that enables data and application portability between the two systems, but they remain separate

entities. For example, sensitive data may remain on the private cloud, but the applications or

services reside on an external public cloud.

The reason for using these specific definitions was to differentiate between internally and externally

hosted services in order to better understand the way optimization solutions are being used and

deployed. Several key elements drove this research study. First, we wanted to verify both the types of

workloads that are being moved to the cloud and the types of clouds in which they are typically being

EMA Research Report: Optimizing the Network

for Reliable Application Delivery Across the Cloud

deployed. Second, we wanted to look at network connectivity for each type of cloud deployment

to better understand how customers are reaching cloud services. Finally, the survey was targeted at

participants who had active cloud deployments across the various cloud types and who had experienced

significant performance issues that impacted end users after initial deployment.

Demographics

This report targeted U.S.-based companies across a wide variety of industry verticals. The participant

base was focused on those in IT organizations and other employees with IT-related roles; therefore,

91% of participants had an IT-related job function, with 9% holding a non-IT executive or managerial

position. Strong emphasis was placed on IT decision makers. Consequently, 60% of participants held

the title of “Director of IT” or higher (including vice president, CIO, CTO, or other senior executive

titles) while the remaining 40% comprised a mix of IT functions, including engineers, supervisors,

and managers. The majority of participants (89%) had a direct hands-on role or a direct managerial

oversight, planning, or stakeholder role for cloud computing projects within their organization. Only

10% were involved indirectly at the technical or managerial level.

From a company-size perspective, we included participants only from organizations that had at least 500

employees, to ensure the organizations had sufficient size and complexity to benefit from optimization

technologies. “Small enterprises,” companies with employee populations of 500 to 2499 employees,

made up approximately 36% of the respondents. The “medium enterprises,” with employee populations

in the range of 2500 to 9999, comprised 29% of the respondent base. The “large enterprises,” with

10,000 or more employees, made up 35% of the respondents.

Participants from the following three verticals were the most statistically significant: finance/banking/

insurance (16%), manufacturing (14%), and high tech (17%). Professional services (9%), education

(7%), government (7%), and retail/wholesale/distribution (7%) were slightly less significant. The

remaining verticals (22%) were grouped for comparison purposes.

EMA sought a survey sample that spanned the three types of cloud deployments as evenly as possible,

while understanding that the rate of hybrid cloud adoption still lags slightly behind rates for public

and private cloud adoption. Individual participants were able to select as many of the different types of

cloud deployments they were using, and the sample was split as follows: private (39%), public (37%),

and hybrid (25%).

Cloud Adoption in the Enterprise

Cloud Services and Adoption

As barriers to cloud adoption decrease, cloud computing is rapidly becoming a core component of the

overall business model. An overwhelming majority (73%) of participants indicated that cloud computing

was either an “essential” or an “important” part of their business, and only 8% said cloud computing

was just a “supplemental” component. (See Figure 1.) As is often the case, smaller companies tend to

more rapidly adopt newer technology while larger organizations, due to their size and geographic reach,

are slower to respond to technology shifts. Small enterprise shows an even stronger emphasis on cloud

with 47% of respondents from smaller companies seeing it as an essential component of their business.

While medium enterprises see it as important to their business (45%), only 30% see it as essential. In

the case of large enterprises, 34% see it as essential and only 25% see it as important.