Organizational Challenges in Hybrid Cloud Adoption

Share this post:

An Infrastructure Perspective

Florin Manaila, Senior IT Architect and Inventor, IBM Systems Hardware Europe

Five years ago, when companies were building their IT infrastructure, nobody was asking about API availability within components such as servers, hypervisors, storage, network, etc. Nowadays, wherever I go, every discussion about IT infrastructure is focused on the following questions:

• How can an IT infrastructure component be integrated into the data center management plan?
• How can the multitenancy be implemented?
• With what infrastructure API’s I need to provide my organization?
• How can the infrastructure be integrated and orchestrated within the cloud to automate organization workloads (ex. marketing campaigns, IoT, mobility, etc)

Many of these questions are new and people are not clear why these questions are becoming a hot topic right now and how by solving the “puzzle” organizations can improve existing IT infrastructure to support new demands in today’s business world.
.

The Puzzle

Everyone is aware of the following benefits and advantages of Hybrid Computing:

• Cost saving – One of the benefits of implementing a hybrid cloud is not having to spend money and to build infrastructure that can withstand occasional bursts in the system usage which only happens on rare occasions. Instead, organizations can leverage public cloud offerings to offload some of the heavy usage and therefore pay for it when they need it. Additionally, a hybrid cloud offers the opportunity to put the development systems and disaster recovery infrastructure into the cloud at a much lower cost.
• Improved security – While many members of traditional IT teams perceive the cloud as insecure, many believe that customers of service-provider environments actually suffer from fewer attacks than on-premise users. The myth that the cloud computing is less secure than traditional approaches stems from the belief that having things stored off-premises is less secure, which is not true.
• Enhanced organizational agility – By leveraging the public cloud in times of heavy workloads, organizations can experience fewer outages and less downtime. The hybrid cloud hosting solution offers an attractive option for developing and testing new applications, which enables organizations to buy some time until they make a decision about the location of the permanently hosted application.
• Greater accessibility – With employees becoming increasingly mobile, greater accessibility to business-critical applications is a necessity for any 21st century enterprise. The days when employees needed access to their email, updated a spreadsheet or accessed an application only during business hours are long gone. Today’s business world operates 24/7 and cloud computing allows the companies to compete effectively as the cloud can be accessed from anywhere and at any time.

However, I find it unlikely that enterprises would follow this direction beyond the use of email and collaboration services. This is not because choosing the right technology and related cloud services is very difficult for them since they may find the following concepts intimidating:

• Compute Virtualization
• Network Virtualization
• Software Defined Storage (SDS)
• Data Center Management Plane

In addition, choosing from a multitude of software components that one should infuse in their datacenter infrastructure is difficult, given the frequency of today’s software release cycles. For example, OpenStack project teams produce a large variety of code repositories. Some are services that provide infrastructure APIs. Some are libraries being consumed by those services. Some are supporting cast and tools. So, typically, OpenStack is released twice a year and every release adds new functionalities or architectural changes.

It is very clear that every customer would like to be an independent vendor and embrace open standards that would give them the freedom of choice and speed. But there is more than that! They want simplicity and a possibility to move enterprise workloads or parts of them back and forth or closer to users in order to achieve competitive advantage.

“The greatest enemy of knowledge is not ignorance; it is the illusion of knowledge.”– Stephen Hawking

I would like to give an example of a company that has decided to implement a mobile payment platform, investing over €2 million in software and hardware infrastructure and spending over a year in planning, designing, developing, testing and implementing. This is one of my favorite examples because it was a successful project for the IT Department, yet a disaster for the Alternative-Channels Department. This mobile payment platform aimed to:

• Add more new customers, being considered as one of the first in the region
• Add new transactions to ease the use of mobile devices
• Sell additional services

After six months of operating the platform, the company gained around 114,000 users – 5% of them were active on a daily basis, while the average user transaction was below €60; the net profit from these transactions was scarce.

They were not wrong in approaching the market with a mobile application but they were too late. Another organization started a similar project two months later, but they took a different path by connecting the existing systems (SoR) to a mobile backend infrastructure in the cloud that gave them the possibility to finish the project within seven months and with a very small budget compared to the €2 million project of their competitors.

Partnering for Hybrid Cloud Transformation

If you are already considering cloud as a sales platform or a cloud solution on its own, you would have come across hybrid cloud. By now you know it is a powerful combination of technologies providing your organization with the maximum scalability, lightning-fast transaction response times and Enterprise DevOps tools for agile development.
The current enterprise infrastructures allow administration through a web-based Graphical User Interface (GUI), Command-Line Interface (CLI) or Storage Management Initiative Specification (SMI-S). However, this is not sufficient in order to move to a hybrid cloud computing model as such a move requires more administration and leads to many operational issues. In such cases, RESTful APIs (related to Representational State Transfer (REST) Application Programming Interfaces) offer a better option.

Over the last few years, REST, which represents a simpler alternative to SOAP, has emerged as a predominant web service design model. The RESTful application programming interface (API) establishes a mapping between the create, read, update, and delete operations and the corresponding POST, GET, PUT, and DELETE HTTP actions.

The RESTful interface focuses on the components’ roles and resources and ignores their internal implementation details. The RESTful API is a client/server model. Therefore, the requests are sent to a server that is component-aware, which masks the intricate details from users.
For example, in a storage environment, a GET request to volumes lists all volumes, and a DELETE /volumes/1 request deletes the volume with the ID 1.

In order to do so, we need to transform current infrastructure and be able to easily orchestrate and automate everything.

Storage systems should be able to offload data into the cloud based on object storage technology for long-term retention (archiving) or for operations such as snapshots and clones. In the economy of speed, the storage tiering-based architecture is now moving to all flash-based arrays and object storage enabled by SDS.

The management of platforms in the compute virtualization should offer OpenStack APIs and the possibility to orchestrate them with infrastructure patterns to support the business processes – that is the best approach for IT transparency in business. In order to achieve this, the OpenStack Cloud Controller and the Orchestrator distribution/vendor must be chosen wisely especially in terms of long-term support, integration, and innovation.

Cloud Orchestrator must offer APIs to support DevOps and integration with physical infrastructure (servers, network, storage) for bare-metal provisioning as well as Cloud Brokerage capabilities. Most enterprises follow the strategy of working with at least two cloud providers.

Conclusion

Yes, infrastructure does matter when it comes to hybrid cloud as both, data and the processing part have to reside somewhere – they cannot just float in the air. It is important not to forget that! But it is the infrastructure that brings superior business value while reducing operational costs.

In addition, the hybrid cloud does not only refer to on-site and off-site workloads. It offers an opportunity to manage infrastructure with the existing cloud-based products. This is what you get, for example, in terms of storage and backup software:

• Capacity forecasts, capacity growth based on empirical data, such as historical growth rates and available capacity. It provides a detailed view of device capacity utilization, thin provisioning, and compression savings.
• Performance displays an integrated view of all the storage systems performance with drill-down capabilities to facilitate performance-related troubleshooting.
• Reclamation identifies provisioned but unused capacity.
• Tier Planning helps to optimize data placement by providing retiering? recommendations.

The advantage here is that such a service is managed from the cloud, getting automatically updated and equipped with new features. This is a very secure way because there are mere recommendations and insights of the infrastructural environment received, without the possibility of changing the operational aspects of the infrastructure.

The same model can be applied to virtualization, extending the existing enterprise capabilities from virtual machines (based on ESXi, KVM, Hyper-V, PowerVM) to use containers and run Docker containers in a hosted cloud environment. A container has the benefits of resource isolation and allocation but is more portable and efficient than, for example, a virtual machine. This results in many benefits, such as building high-quality apps fast, no maintenance of the required infrastructure for updates, securing the docker etc., and all delivered automatically by the cloud provider.

In conclusion, it is essential for any organization to choose the right partners in this transformational journey which is far from easy. Many cloud projects failed because they were led by:

• an organization that lacked skills and products in terms of hardware infrastructure (compute, network and storage),
• an organization with no experience in legacy applications, or
• an organization that had no knowledge of new operational aspects introduced by the cloud computing model and how such aspects influence the existing enterprise operations.

What’s next?

I have been asked many times: “What’s next?” Now, that everything is stored in one place, almost everything is automated, thinking about the next step is mandatory if the company want to set its strategy right. The majority of enterprises have built their IT infrastructure and operation around structured data but, for the IT, unstructured data is the new frontier offering tremendous opportunities and endless possibilities for innovation. The IT infrastructure will have to change the mindset of data storage, processing and analysis, “learn” from data (supervised learning or unsupervised learning), and leverage hardware acceleration systems. In other words, this means that, in the long run, everything related to structured data will be considered history and companies need to have their IT infrastructure ready for that shift.

Do you have questions?

Come and join me on September 7th, in Zagreb at the SEE Cloud Computing Forum to further discuss Hybrid Cloud solutions and ways you can use them to transform your business.
I will be speaking about cloud services and how they allow companies to innovate iteratively with less investment.
I will also discuss IBM’s point of view on that which outlines the spectrum of IBM capabilities – from private to hybrid to public cloud – offering clients flexibility and choice to deploy in a way that is best suited for their business. What differentiates IBM in that journey, is that we help clients leverage cloud with confidence, providing the security underpinnings at every touch point.