Cloud/SaaS/SDN, Forum

Part II – Building and Managing Enterprise Cloud Systems – Deploying Hybrid Cloud Challenges

In the previous post, we covered the challenges of building, managing and scaling cloud architectures from an orchestration and automation point of view.  In this post, we explore one other challenge in deploying cloud architecture – deploying hybrid clouds. The business drivers for cloud deployment are strong enough to continue the push towards cloud infrastructure. While it is easy and quick to build infrastructure on public cloud offerings due to quick IT provisioning, multiple drivers create the need for an enterprise to build a hybrid cloud – a combination of public and private cloud infrastructure. First, enterprises in Financial and Healthcare verticals tend to be far slower than other enterprises to move towards a complete public cloud deployment due to a variety of regulatory and privacy constraints. For enterprises in these sectors only non-critical frontend infrastructure can be moved to the cloud. Furthermore, most enterprises have invested in infrastructure that represents sunk costs for them. Business managers are under ever increasing pressure to extract value out of these sunk investments. Finally while building and scaling public cloud and data centers is relatively easy, the spending on this public infrastructure can grow very quickly. All of these factors create a push towards the adoption of hybrid cloud infrastructures. Building a private cloud has its own set of unique challenges. This includes specialized IT resources with the necessary skill sets as well as the longer times required to provision private cloud services. Software such as Eucalyptus, Openstack and CloudStack enables companies to build private cloud infrastructure similar to the public cloud themselves. Enterprise software vendors facing requests from their customers to build out a private cloud infrastructure can leverage these tools to do so as well. The growth of hybrid architectures is beneficial to enterprises since it provides them the flexibility to store and access data in locations that they are comfortable with. However, such architectures often create cloud information silos. Cloud computing is more than just a fast self-service of virtual infrastructure. It provides an opportunity to centrally analyze data and extract meaningful insights from data that can a) improve enterprise productivity, b) empower fast and data-backed decision-making and c) decrease costs. But with hybrid cloud architectures, information is now spread across multiple systems, making it harder to have a “single version of truth”. While several business intelligence and SIEM tools provide cloud and enterprise offerings, there is a gap in products that offer tools to seamlessly combine information from both public and private cloud systems and offer a “single version of truth” for the business user. This is an area for innovation. We believe that, while cloud computing technology has matured, it will continue to see exciting innovation along multiple dimensions, with startups providing newer ways for enterprise data management and storage, or startups exploring newer business models to gain market penetration. An example of product innovation is FormationDS – the company is working on a ground-up architected data services platform that aims to transform a traditional client-server computing model to hyper-scale computing, primarily driven by modern application models that seek to compose capabilities as API-based services. This is a pure object oriented approach to computing. On the business model innovation front, we see an increasing use of opensourcing of software to tackle market penetration challenges within the cloud management space. OpenStack and Scalr are perfect examples for this type of innovation. While making purchase decisions, enterprises face a big supplier risk, especially while dealing with new software vendors. As such, enterprises are left wondering if the new businesses will be active a few years from now. In fact, a larger vendor can acquire these companies and force them to align with a particular hardware vendor or architecture. Making the software open source alleviates this risk to enterprises, since the enterprises have access to original software under all circumstances, thereby opening up opportunities that would otherwise be closed.