INDUSTRY PERSPECTIVE / THE DIGITAL EDGE
BY JOHN YUNG, FOUNDER & CEO, APPCARA INC.
PUBLISHED 30 JULY 2016
The Cloud journey has been challenging for most enterprises as technology evolves rapidly with vendors frequently repositioning themselves. What seems to be the right choices today can be very different in a year. As a result, enterprises commonly own and operate multiple Cloud technologies. This can be very inefficient and costly.
The Cloud phenomenon has been disrupting enterprise IT broadly and deeply in a short amount of time. CIOs must apply practical strategies to transform from traditional on-premise to hybrid IT environment. The hybrid IT platform must have the flexibility to integrate various technologies and provide a unified control panel for managing the entire environment including applications.
Cloud Started Simple but Not for Long
Cloud started a decade ago simply to offer virtual servers for on-demand and pay-as-you-go consumption. The ability to deploy servers in minutes and not needing to spend capex was disruptive. The trend has quickly evolved into public, private and hybrid Cloud models. Enterprises started adopting public Cloud for non-production workloads and private Cloud for workloads requiring predictable SLA and performance.
Although functionalities among these Cloud technologies are similar, user experience can be vastly different as well as underlining formats and programmatic interfaces (API). The lack of portability and interoperability between these technologies cause vendor lock-in concerns. The introduction of OpenStack and recent market consolidation make choosing Cloud platforms more challenging. Open source Cloud orchestrators are free but require deep technical expertise to implement and operate while commercial vendors are constantly realigning their products.
Implementing a future-proof hybrid Cloud platform is a difficult task because of the broad and deep technical requirements. Cloud involves the entire infrastructure stack including server, network, storage, hypervisor, Cloud orchestrator, firewall and load-balancer – just to name a few.
One of the challenges is to size the private Cloud environment appropriately. It requires accurate calculation of physical processor cores and memory (RAM) as well as good understanding the applications intended to be run. Example: Big data applications may require 8+ vCPUs and 32GB+ vRAM to run efficiency. Choosing low-end processors with less than 8 processor cores or less than 32GB of physical memory on the host can be limiting.
According to Gartner, hybrid Cloud will be the predominate enterprise deployment model for the foreseeable future. It offers enterprises the flexibility of usage base billing on public Cloud and fixed cost on private Cloud. Web applications with seasonal traffic is an excellent fit for the hybrid Cloud deployment model. But CIOs must realize the added complexity.
Augmented Reality games (i.e. Pokémon Go) are excellent applications to leverage Cloud for delivery and scaling. For the entire application stack to work properly at scale, automation plays a key role. The network, server and application must scale up and down together. Having a unified control panel is key element for success.
Apps Heading to Cloud
Ultimately CIOs want to operate their apps in the Cloud so that legacy environments can be retired. After a decade, enterprises have varying degree of successes. Similar to Cloud technology, apps have different deployment models as well – hosted and on-premise. Also, certain legacy apps don’t function well in the Cloud. CIOs have to carefully choose the right model for their organizations. Modern enterprises typically have mixed SaaS and on-premise app environments.
Salesforce started in 1999 with the vision for enterprises to consume applications in a hosted or “SaaS” model. Its success accelerates other SaaS app adoptions such as Microsoft Office 365 and PayDay for enterprises. SaaS is perfectly suited for apps with standard workflows and usability.
But enterprises also have apps that require ongoing development and enhancements such as e-commerce sites and ERP systems. These apps aren’t suitable for the SaaS model and therefore usually operate either on-premise or in commercial data centers. Many of these on-premise apps are migrating to the Cloud. As a result, most enterprises are operating a mixture of on-premise and Cloud app environments.
Automation plays an important role in future-proofing enterprise apps. It’s impractical to manually provision various app components as load increases. These app components must be pre-packaged with intelligence on their relationships and configurations. Example: Apps rely on various SQL and non-SQL databases for user profiling and keeping game states. Once the policy-based scaling engine determines more web servers is required, the provisioning engine would dynamically calculate the number of web and app servers required as well as database instances and deploy them in locations needing the extra capacity. The entire provisioning process must be automated.
Automation plays an important role in future-proofing enterprise apps
Most enterprises today are operating more complicated IT environments than the pre-Cloud era with mixed bare-metal server, virtualized and hybrid Cloud environments as well as on-premise, SaaS and Cloud apps. Each of these components usually have its own control panel and support arrangement.
To operate today’s IT environment, enterprises are required to hire top talents, train them on various control panels and keep up with new technology releases. There is a big shortage of skilled IT administrators with broad skill to design, implement and operate these technologies. Although most enterprises aren’t nearly at the Pokémon Go scale, it’s important to adapt similar practices. CIOs are eagerly seeking single-pane-of-glass control solutions for their entire Cloud environment to reduce complexity and gain efficiency.