For technologically tuned-in senior executives, the potential benefits that can be gained by making the move to cloud computing may appear to be nothing short of spectacular. After all, there is a lot of hype around the promise of benefits such as fast time to implementation, cheap infrastructure in a pay-as-you go model, substantially reduced electricity and overall IT costs, and plenty more.
But when CEOs and CIOs take the time to talk to their IT departments, they are surprised at the number of new complexities introduced by cloud computing. Software and cloud service vendors must understand the real-world business impact of the cloud-based delivery of IT services in order to help make these deployments successful at client organizations over the long term.
Business challenges of cloud computing
To begin, I want to clearly define what we at UC4 define as a cloud computing. In order to qualify as a true cloud deployment, we believe these five elements must be present.
1. Multi-tenancy – Users share applications on a set of pooled hardware resources
2. Virtualization – Infrastructure is virtualized between clients
3. Service-based – Service levels are established and vendors commit to them
4. Scalable demand – Application and infrastructure is tightly integrated so that it expands and contracts based on usage needs
5. Adjustable billing – Charges are metered based on usage levels
This checklist helps clients understand whether a public or private cloud they are planning is really indeed a cloud – or only virtual infrastructure.
These five cloud characteristics also provide a clear context for the importance and relevance of the new business challenges which arise from cloud-based delivery of IT services in three key areas: service levels, decision automation and system integration.
1. Managing the complex new world of internal service levels
In the past, IT was viewed as a cost center or administrative arm. The department would charge a line-of-business (LOB) group to install a new application and for the servers it needed to run it.
In the world of the cloud, IT is delivering systems via some type of defined service level. This changes both how IT is viewed internally and how its services are consumed.
But a line-of-business manager doesn’t care about traditional technology-based service levels – exactly how many servers are powering a particular application or where it is hosted. He or she will only be concerned with service levels expressed in business times: X% of uptime, X second-maximum response time, and so on.
The need for business-specific service levels presents a new internal challenge from both a business and a technical standpoint. On the business side, SLAs between business groups and IT will involve a new language of business-based terms and will move away from the technical SLA language historically used by IT with its external vendors and internal customers.
On the technical side, any application running on external IT infrastructure will still need to be managed by internal systems in order to maintain guaranteed service levels. As companies install their first cloud apps and establish their first two or three internal service level agreements, it is possible to try to manage this manually. After that point, however, the complexity increases exponentially and service levels must be automated in order to maintain control and to continue to expand service as needed.
2. Automation of the entire decision process
In IT’s pre-cloud days, system management and administration could be handled by department personnel. When only a finite number of applications were involved, servers could be provisioned and functions such as load balancing could happen periodically when needed.
In a cloud-driven IT environment, virtual sessions are spun up and down constantly, apps are engaged then idle, users are online then offline. Change is constant.
I like to think of this evolution like the one that happened with air traffic control. In the days of the Wright brothers, there was no need for such a system because there was so little traffic. Today, the skies are littered with aircraft, weather patterns and competing priorities – not unlike the cloud computing which powers many enterprises.
Today, IT needs to automate its entire decision making process. Organizations must not only decide which workload will run where – public, private or hybrid – but automate all the decisions around that in order to maintain availability and response time. The process needs to be the same as if a car could automatically turn on its wipers when a raindrop hits the windshield.
3. Consideration of the entire enterprise computing ecosystem
Early adopters have often set up their cloud-based systems separate from their legacy systems. These CIOs aim to set up a pristine cloud environment, deciding which workloads and applications can run in the cloud and which need to stay in-house.
This silo-ed philosophy is short sighted. Embarking on a cloud computing strategy is not just about setting up a new and different platform; it is about evolving the overall IT strategy to be the most efficient environment possible.
The key is integrating all computing environments. By building bridges from public and private clouds to legacy systems, it is possible to set up a scenario where handoffs and exchanges of workloads between applications are handled automatically without losing data or files. These transactions must be handled seamlessly between all the layers and automated in order to operate successfully on a long-term basis in the cloud.
Fundamental changes in IT
The new challenges presented to IT by cloud-based services are changing the face of IT within the company and the nature of its relationship with its users.
To begin, IT needs to set new expectations with end users and endeavor to meet the heightened expectations users will have of IT. This will require a new dialogue and a new mindset on the part of IT and business managers, alike.
Over time, the role of IT overall will evolve from head-down, task-oriented administrators to trouble-shooting, strategic planners. As more of the manual jobs and routine decisions are automated, IT staff and managers must become business experts who are creative at using technology to improve productivity, enhance the user experience, and/or improve the bottom line.
And as more computing moves to the cloud, IT must become more efficient on its own. In pre-cloud IT days, a LOB manager unhappy with the service provided by his or her IT department would have to begin a months-long process to outsource the workload. In the cloud, IT departments are competing with Amazon Web Services: if IT cannot deliver, an LOB manger could move a system to a third-party service in the cloud in just a few hours.
Implications for software vendors
Software and other types of cloud vendors can either help companies and their IT departments make this cloud-driven business transformation – or they can hold them back. In the field today, I have seen a few ways in which software companies are not acting in the best interest of their customers:
- Encouraging cloud silos” – There are many virtualization companies which advocate that enterprises do not integrate their legacy or physical-node environments with their cloud systems. Other vendors tell clients it will be 10 years before they can move their legacy systems. By encouraging customers to set up walled environments, they are “hamstringing” the success of their overall IT strategy in the future.
- Tech-spec SLAs - Some cloud vendors have begun to advise clients about internal service levels – but only infrastructure service levels, with metrics like CPU utilization and bandwidth. But a user of a SAP application doesn’t care what the uptime is for the server; he or she wants to receive a report within 10 seconds. Service levels expressed in terms of speed, usability and response are what matters today.
- All-or-nothing SaaS clouds – Many SaaS application providers want their clients to put everything in the vendor’s data center in the cloud. But this decision should not be driven by application – it should be driven by workload. Real-world companies operate in complex environments governed by different regulations in different regions: HIPAA requires certain data to reside in a private cloud in a VPN, certain French companies can’t have their data outside of France, and so on. Compliance requires flexibility – not an all-or-nothing mentality.
Cloud computing is delivering significant benefits to the leading-edge companies that have adopted the technology to date. In order for these and other companies to realize the true, future potential of the cloud, software vendors need to help IT departments adjust not only to the cloud’s technical implications, but to its business and organizational impacts as well.
Jason Liu is CEO of UC4.