Until recently startups and small to medium businesses drove cloud adoption. Now, with cloud service providers maturing, many businesses are considering migrating applications to the cloud. In particular, the IT industry has witnessed a rapid adoption of the public cloud. This research report examines the current state of the cloud, barriers to enterprise cloud adoption, and the key factors to consider when choosing an enterprise cloud platform.
Since the majority of cloud-based applications and workloads require elasticity, the major public cloud adoption trend is toward hosting scale-out applications like a public-facing web application or a mobile backend that exposes API. Enterprise customers usually begin evaluating the cloud by moving development and test environments. They have yet to broadly migrate mission-critical applications like ERP, CRM, messaging, collaboration, business intelligence, and other line-of-business applications.
Many factors influence the decision to migrate enterprise applications to the cloud. Many applications must remain in the data center while other applications move to the cloud. This division requires a robust hybrid platform that can seamlessly bridge the cloud and on-premise assets. Reliability, security, availability, compliance, and governance all play a major role in the decision.
Cloud providers must supply their customers with an enterprise-ready platform. Enterprise applications need an environment that is reliable and robust. It should deliver predictable performance. Not every cloud platform is enterprise-ready.
Key findings in this report:
- The ROI of cloud migration varies by workload, those with potential gain being most ready. Enterprises can start cloud migration with these cloud-ready workloads and then move mission-critical applications.
- The cloud addresses discrepancies between dev/test and production by automating the provisioning environment that closely resembles the production environment.
- Despite understanding the benefits of the cloud, CIOs will not move their core business applications into the cloud until cloud providers address several main concerns, including security, control, customization, and complexity.
- Enterprises prefer integrated, managed cloud service offerings that provide high availability, security, compliance, and assured business continuity.
2 The current state of the cloud
Though it started as a phenomenon primarily driven by startups, the public cloud is gaining traction among all customer segments, including large enterprises.
The available Infrastructure-as-a-Service (IaaS) offerings in the market are designed to run a diverse set of workloads ranging from public-facing marketing websites to an ERP application. The characteristics of each workload are distinctly different. The one-size-fits-all approach doesn’t make the public cloud an optimal environment for running enterprise applications. Most of the applications running on generic IaaS platforms are designed to take advantage of cloud elasticity. They can easily scale out to run across a fleet of inexpensive, low-end virtual servers. Traditional enterprise line-of-business applications are not elastic, and they perform better when configured to scale up hardware resources rather scaling out. Increasing the number of cores, primary memory, and storage will have a significant impact on the performance of enterprise applications.
The other major challenge that enterprise IT teams face when dealing with generic IaaS platforms is the networking stack. Though there have been many enhancements brought to the networking layer of IaaS, emulating the exact topology and configuration of a corporate network on the public cloud is difficult. This disparity in the network architecture widens the gap between on-premise and cloud deployments.
No enterprise follows an all-or-nothing approach when considering the cloud. Customers are looking at retaining mission-critical assets within their data centers while migrating some customer-facing and partner-facing applications to the cloud. They want a well-defined hybrid strategy that seamlessly connects corporate assets with the cloud. During the past year, the hybrid cloud gained traction with support for diverse enterprise scenarios. However, hybrid architectures supported by generic cloud platforms still lack maturity to support the full spectrum of enterprise IT scenarios.
Most of the available public cloud platforms cater to developers rather than enterprises. By exposing APIs and SDKs, public clouds make it easy for the technical community to consume their services. However, CIOs find it hard to relate with these offerings, as they fail to articulate a clear enterprise value proposition. It may be easy to deploy a Ruby or a node.js web application on these platforms, but they lack the right documentation, evaluation guides, and best practices for migrating enterprise workloads. Little to no guidance is available for migrating enterprise applications to these platforms.
3 Common enterprise workloads
The ROI from cloud migration varies from workload to workload. Cloud can be a game changer only for certain workloads. Cloud-ready workloads are those that are aligned in terms of their potential gain and ease of deployment. These workloads need not be customized, because they take advantage of the benefits offered by the cloud. Enterprises may start the cloud migration with cloud-ready workloads before attempting to move mission-critical applications.
Following is a list of common enterprise workloads well-suited to migrating to the cloud.
Almost every enterprise starts its cloud journey by moving its development and test (dev/test) environment to the cloud. When development teams request resources for new dev/test environments, it may take weeks for the IT team to procure and provision the specified hardware for dev/test needs. Furthermore, the typical utilization of these dev/test servers is less than 30 percent. The development and quality assurance (QA) teams need an environment that reflects the production environment. For the IT departments, it is a challenge to emulate the production environment within the dev/test setup. The discrepancies between dev/test and production result in the slower development, building, testing, and deployment of software. The cloud addresses these issues by automating the provisioning of the environment that closely resembles the production environment. The servers within the development environment that are not being utilized can be terminated and brought up only when required, resulting in cost benefits.
Public-facing production websites
Enterprises prefer to migrate workloads that are stand-alone with no dependencies on the internal systems. Public-facing websites are therefore ideal candidates for the cloud. These websites may be deployed as part of a marketing campaign; they are short-lived, but they experience high traffic. The data captured by these sites can be asynchronously merged with backend systems. By running this workload on the cloud, customers take advantage of elasticity and automation. Resources can be added or removed as needed, based on the traffic patterns.
Collaboration platforms connect internal and external stakeholders. For internal users, these platforms increase collaboration and productivity through a centralized repository of documents and resources. Externally, collaboration platforms enable organizations to locate and stay connected with the customers to keep them informed on the latest products and services. By running collaboration platforms on the cloud, enterprises are able to reduce the cost of managing applications. Shared repositories of data allow both internal and external users to search and retrieve relevant information more easily.
Data warehousing as well as business continuity (BC) and disaster recovery (DR) are two use cases that are particularly well-suited to moving databases to the cloud.
Enterprises start by moving OLAP data stores to the cloud. With transactional systems running on-premise, the extract, transform, and load (ETL) jobs are completed and the final data sets powering analytics are moved to the cloud. By running the data warehouse and the visualization and reporting tools in the cloud, enterprise customers are able to save on the cost of compute and storage to host the OLAP application.
BC and DR in the cloud is another common scenario preferred by many enterprises. With the cloud as the DR site, it is easy to meet the desired recovery-point-objective metrics and recovery-time-objective metrics defined by the business. DR in the cloud offers significant cost savings to enterprises.
Email is the backbone of enterprise communication, demanding high availability and uptime. By moving messaging servers to the cloud, enterprises can adjust to fluctuating requirements by adding compute and storage dynamically on demand and balancing automation and DR planning to achieve the required uptime. Enterprise messaging-software providers such as Microsoft and IBM offer license mobility that brings portability to the software.
Enterprise resource planning
Running enterprise resource planning (ERP) in the cloud enables dynamic scaling and disaster recovery to enterprise customers. Many customers are extending SAP to the cloud for using it in production and non-production scenarios including testing, training, sandbox, and disaster recovery. When the load increases, it is easy to scale up the ERP deployment with minimal downtime. The database that powers the ERP application can be configured for disaster recovery to avoid prolonged downtime. Enterprises are finding it easy to manage various ERP environments related to development, testing, staging, and production on the cloud. Finally, some of the reports indicate that running ERP in the cloud offers better TCO than running it on premise.
4 Enterprise cloud-adoption barriers
In 2013 KPMG conducted a survey asking over 600 enterprises on the key challenges to adopting the cloud. Of the participants, 33 percent responded that the implementation and integration costs are too high, while 31 percent responded that integration with existing architecture is a challenge.
Though cloud computing has been on the agenda of enterprise CIOs, they remain cautious in migrating critical workloads to the cloud. While CIOs understand and appreciate the benefits of the cloud, they will not move their core business applications into the cloud until cloud providers address eight main concerns. Enterprises should evaluate how these factors impact their cloud-migration strategy.
- Security. Security is the biggest concern of enterprise customers. Cloud service providers need to implement the highest level of security controls, ranging from physical security of data centers to the encryption of the data stored in the cloud. Current industry certifications like ISO 27001, SSAE 16 Type II, and PCI DSS ensure that the right controls are in place.
- Control. Enterprise IT fears lack of control when dealing with the cloud. Each organization has a set of policies to ensure smooth IT operations. This includes everything from locked-down desktops to data-retention policies. Current cloud platforms do not offer the level of granular control that enterprise IT environments demand.
- Customization. Many cloud environments are opaque for customers. For example, mapping the on-premise hardware specifications to the virtual server configuration offered by self-service cloud providers is difficult. Enterprise customers don’t appreciate the straightjacketed approach of the cloud service providers. They need customization at various levels to ensure alignment with corporate environments.
- Complexity. The complexity of public cloud providers acts as a barrier to enterprise adoption. Many customers find that taking the first step to the cloud is difficult because of the cloud’s confusing and complex choices. Emulating existing network topologies on the cloud and implementing a hybrid strategy is complex and challenging for enterprise IT teams.
- Cost. The current pricing model of public cloud providers is confusing. To arrive at a simple TCO calculation, customers have to deal with ambiguous parameters like the amount of inbound and outbound network traffic and the number of I/O operations performed on cloud storage. Though the prices of compute and storage are dropping almost every quarter, enterprise customers are not convinced that the cloud is cheaper than running a data center. The debate of capex versus opex does not apply to enterprise customers, who typically have million-dollar annual IT budgets. Cloud service providers should enable enterprise customers to calculate the TCO of migrating to the cloud easily.
- Confidence. The confidence of running enterprise workloads in the cloud is low among CIOs. Cloud service providers must be transparent in sharing how they achieve and maintain compliance. For example, health-care-related companies are looking for HIPAA compliance before they make a move to the cloud. Various standards like SOC 1, SOC 2, SOC 3, FIPS 140-2, ITAR, and FISMA provide the right level of confidence to enterprise customers.
- Technical skill set. Lack of a required skill set within IT teams is slowing down cloud adoption, particularly because of an acute shortage of skilled professionals who can deal with cloud-related technologies. Monitoring, managing, and optimizing cloud-based deployments is fundamentally different from managing workloads running in traditional data centers. Though the engine that powers most of the cloud infrastructure is virtualization, public and hybrid clouds demand a different set of skills. Few industry certifications endorse the skills of IT professionals for performing cloud-specific operations. Cloud service providers should invest in training and certification programs to enable the large IT workforce to embrace the cloud.
- Tools and processes. The tools required for managing assets across on-premise, private clouds and public clouds are fragmented and thus costly for enterprises. Existing processes and frameworks based on ITIL, ITSM, and TOGAF do not align with cloud operations. The fragmentation of tools and lack of operations framework are slowing down the adoption of the cloud among enterprise customers.
5 Key factors to consider when choosing an enterprise cloud platform
Almost all the self-service public clouds today try to be everything to everyone. They want to attract hobbyists, developers, startups, and large enterprises. But enterprises have different needs from independent developers and startups. They need optimized hardware, proven virtualization technology, and a high-performance cloud orchestration engine.
The hybrid cloud is the biggest enabler of the enterprise. Following are four key factors that differentiate enterprise cloud platforms from generic public cloud offerings.
Consistent private, public, and hybrid cloud environments
Private clouds are witnessing more increased adoption than public clouds. According to a survey conducted by InformationWeek in December 2013, private cloud adoption has increased from 21 percent to 47 percent in just one year. Enterprises with significant investments in a private cloud should look for a public cloud that is consistent with their private cloud implementation.
If the private cloud and public cloud are based on the same hypervisor, cloud-orchestration software, cloud management, and governance models, customers gain more efficiency in managing the environment. They deliver the same user experience to the IT staff dealing with the life cycle of applications and servers. Consistent APIs enable easy integration with existing applications and tools. Since the underlying networking stack is the same, maintaining the same level of isolation and security across private and public cloud environments is possible. This integrated environment enables the virtual private cloud that combines enterprise-class attributes with the full public cloud. Cloud providers like Rackspace, CenturyLink, and CSC have offerings that cover private, public, and hybrid clouds. Dimension Data’s Managed Cloud Platform (MCP) is another example. Its Public MCP, Private MCP, and Hosted Private MCP offer consistent performance and manageability of the platform.
Source: Dimension Data/Gigaom Research
Enterprises considering moving to the cloud should look for partners to help them with migration and system integration. Cloud service providers offering professional services are preferred for their knowledge and experience dealing with platforms. They work with customers to architect solutions that integrate legacy environment with the cloud platform. As a part of the transformation, they identify the required changes to the infrastructure — from network, communications, and security to the data center, end-user computing, applications, and service management — to ensure the new service delivers the desired return on investment and business outcomes. This results in the tighter integration of private, public, and hybrid cloud environments. With cloud providers delivering professional services, enterprises get best-in-class performance and cost optimization.
Going beyond the basics
Many cloud platforms offer self-service and elasticity, but enterprises need more than that. When running mission-critical workloads on the cloud, customers need a bulletproof backup-restore strategy. Managed service providers (MSPs) and cloud providers with integrated backup offerings deliver more value. An enterprise class backup-restore service should include backing up and restoring systems, files, and applications to a separate infrastructure block, for redundancy. It should provide visibility into backup status, coverage, and success rates. Having a choice of pre-configured backup windows and retention periods offers granular control when defining the backup strategy.
The next critical need of the enterprise cloud platform is disaster recovery (DR). The cloud enables faster disaster recovery of critical enterprise IT systems without incurring the infrastructure expense of a second physical site. Instead of leaving the responsibility of implementing disaster recovery, enterprise cloud providers offer Disaster Recovery as a Service (DRaaS), which conforms to organizational DR strategy with configurable recovery point objective (RPO) and recovery time objective (RTO).
Cloud infrastructure platforms are highly reliable and available. By implementing a high level of redundancy, cloud providers make fail-over of systems seamless and transparent. These encompass critical building blocks of the enterprise platform including compute, storage, and networking that facilitate maximum uptime of mission-critical enterprise environments.
Finally, enterprise cloud providers offer solid security that covers the infrastructure, information, and identity in a multitenant environment. Infrastructure security is implemented through the isolation of the virtual networking stack. This enables provisioning sensitive workloads in the trusted zones of the network while running the public-facing resources in an isolated network. Information security is handled by applying military-grade encryption to the data at rest and data in transit. Identities are safeguarded in the cloud through the access controls that are in line with corporate policies. Enterprise cloud platforms implement single-sign-on (SSO) multifactor authentication to seamlessly extend the identity and access control to the cloud.
Migrating workloads to the cloud is only half the story. Managing deployments and constantly optimizing them for performance and cost is a major part of successful cloud migration. The general-purpose cloud providers today offer basic self-service monitoring tools that do not meet the bar of enterprise IT standards. Mature cloud providers offer comprehensive monitoring that covers the key performance metrics of operating systems, running processes and custom workload-specific metrics that are critical for application health.
Since enterprise cloud providers take ownership of the entire stack, they ensure that software is up to date by applying patches as soon as software vendors release them. Generic general-purpose cloud providers point their customers to the shared responsibility model, where the ownership of patching is left to the customer. If the software is not patched on time, it leads to unwanted security and stability issues. Customers should check for the patching policy of enterprise cloud providers before migrating workloads.
The next important aspect of managed services is configuration management. This includes upgrading software components and reconfiguring them to reflect the latest IT policies. Enterprise cloud providers offer comprehensive configuration management to customers, freeing them from dealing with prolonged downtime of applications. By using proven tools and methodologies, they manage the deployments efficiently while maintaining uptime.
Customers must evaluate the support levels of the cloud provider. Enterprise cloud providers give their customers access to world-class engineers skilled across a variety of technology areas and available across global locations and time zones. They also offer third-party expertise through their partners, software, and networking providers such as Cisco, Microsoft, Blue Coat, Check Point, F5, Juniper, and Riverbed. The support services offered by the enterprise cloud vendors complement the existing IT support teams.
They deliver better service-level management (SLM) through skilled service-delivery teams with tools and applications to analyze the data and incidents and uncover the root cause of the issues. They will also suggest activities to mitigate recurring incidents. Support forms the key decision-making factor when choosing a cloud platform.
Designed for enterprise workloads
Enterprise workloads need an enterprise-class cloud platform. These cloud platforms are mature, stable, and reliable, and they are designed to deliver better SLAs.
Enterprise applications demand highly structured networks with multiple public or private subnets, network-address translation, and special traffic-routing requirements. Enterprise customers require building hybrid cloud solutions where highly sensitive data and processing remain in the private cloud while other workloads are moved to the public cloud. Enterprise CIOs want to run workloads on the private cloud and then burst demand peaks into a cloud provider’s cloud, thus reducing their own spare capacity and increasing the server utilization. These bursting processes need to be transparent to the applications. Workloads that are moved to the external cloud will remain unchanged, with the same IP address as in the private data center, and they will continue to interoperate with the assets running in the private cloud. Through this strategy, data at rest and data in transit are always encrypted.
Enterprise cloud platforms are designed to emulate the on-premise network. IT teams can extend the set of IP addresses, subnets, routing policies, gateways, and firewalls seamlessly to the cloud. This offers granular control on the networking stack, which is essential when dealing with enterprise workloads.
One of the challenges with general-purpose cloud providers is the ability to map the on-premise hardware specifications with the virtual servers that run on the cloud. Even after choosing the recommended instance type that closely resembles the server specification, the performance never matches that of the on-premise hardware. Customers experiment with a variety of instance types before finalizing the right one. These cloud virtual servers running on low-cost, commodity hardware never deliver consistent and predictable performance. This becomes more obvious when running I/O-intensive workloads like OLTP databases and ERP applications. Since the storage architecture is based on a shared, multitenant model, it becomes the bottleneck that seriously limits the performance of the workloads. Enterprise cloud providers overcome this by implementing the right architecture that delivers predictable performance and throughput to the applications. Based on specialized hardware, high-speed fiber networks, and high-speed SSD-based storage technology, enterprise cloud providers meet the expected performance levels of the complex workloads.
Finally, enterprise cloud providers support the scaling up of hardware instead of the scaling out that is an attribute of generic, self-service cloud players. Many enterprise workloads are designed to run on powerful hardware configurations. When performance degradation is observed, additional resources (i.e., CPU, RAM, and hard disk) are added to the servers. Enterprise cloud providers support this mechanism of scaling up the resources on demand with minimal disruption.
6 Key takeaways
Complex enterprise workloads require mature and reliable cloud platforms. Most cloud platforms are designed to run public-facing websites and elastic workloads rather than complex enterprise software like SAP and Oracle. Only a few of the available IaaS offerings in the market today are designed to run enterprise line-of-business applications.
- Not all cloud platforms are enterprise-ready. Enterprise cloud providers support traditional workloads that require optimized environments.
- Development and test is the first choice of enterprises for migrating to the cloud, followed by messaging and collaboration workloads.
- Enterprises are concerned about security, control, cost, compliance, and SLAs when migrating to the cloud.
- Enterprise cloud platforms offer consistent, predictable performance with seamless integration between on-premise assets and cloud deployments.
- Enterprises prefer integrated managed services from cloud service providers.
- Enterprises prefer cloud service providers that offer high availability, security, compliance, and assured business continuity.
7 About Janakiram MSV
Janakiram MSV heads cloud infrastructure services at Aditi Technologies. He was the founder and CTO of Get Cloud Ready Consulting, a niche cloud migration and cloud operations firm that was recently acquired by Aditi Technologies. He is the principal analyst at Janakiram & Associates, where he focuses on research and analysis related to cloud services. Through his speaking, writing, and analysis, he helps businesses take advantage of emerging technologies. Janakiram is one of the first few Microsoft-certified professionals on Windows Azure in India. He is also an AWS-certified solutions architect. Janakiram studies the cloud services landscape for the Gigaom Research analyst network. He is the editor in chief of a popular portal called CloudStory.in, where he analyzes the latest trends from the world of cloud computing. Janakiram is a guest faculty member at the International Institute of Information Technology, Hyderabad (IIIT-H), where he teaches big data, devops, and cloud computing to the students enrolled in the master’s program.
Janakiram has worked at world-class product companies including Microsoft, Amazon Web Services, and Alcatel-Lucent. His last role was with Amazon Web Services as a technology evangelist, where he joined it as the first employee in India. Prior to that, Janakiram spent 10 years at Microsoft, where he was involved in selling, marketing, and evangelizing the Microsoft application platform and tools. When he left Microsoft, he was a cloud architect focused on Windows Azure.
8 About Gigaom Research
Gigaom Research gives you insider access to expert industry insights on emerging markets. Focused on delivering highly relevant and timely research to the people who need it most, our analysis, reports, and original research come from the most respected voices in the industry. Whether you’re beginning to learn about a new market or are an industry insider, Gigaom Research addresses the need for relevant, illuminating insights into the industry’s most dynamic markets.
Visit us at: research.gigaom.com.