With Microsoft suffering another outage with Office 365 potential enterprises must again evaluate the risks involved with going off-premise to the cloud. I will not debate the proposed benefits of going to the Cloud. The magic transformation of CAPX to OPX has to have every CFO pulling out their spreadsheets and CIOs examining the possibilities of elasticity.
These do come at a cost. Increased risk. While the infrastructure behind the cloud maybe resilient –depending upon the architecture used to implement the solution– Internet Access is still a potential single point of failure. There are multiple places in the communications channel where glitches could take down an entire enterprise.
Imagine your gateway is down. On-premise this is not a problem, end users still have access to services behind the firewall. Another possibility your provider is down, again access to all your services is at risk and then in the case of this latest glitch, and unconfirmed network appliance on Microsoft’s premises took the service down for three hours.
This is not an article bashing Azure or Office 365. As an Enterprise Architect I find Cloud and specifically Azure a useful approach. Actually before Cloud was Cloud I was writing articles about Virtual Corporations http://intelarbgrp.org/Documents/Engineering%20a%20New%20Paradigm.pdf using the Internet to create a virtual service infrastructure between corporations -and had demo’d some of these concepts at Autofact and prior to that discussing the digital nervous system at IBM Guide/Share conferences in the 80’s.
But like any tool its only as good as the skilled hands that use it. What I mean by that is when considering a move to the Cloud multiple factors should be considered beyond reducing cost. These factors need to be weighted carefully at the enterprise level and brought to the attention of business executives.
About a decade ago corporations when through a scare called Y2K, may companies spent millions on prevention. One of the large Fortune 500 corporations clients I had the meter spinning with all the contractors they brought in to fix the possible issues. They were concern that they were wasting money by the truckload and asked me to assist. I developed an approach, or rather adapted a systems engineering approach [FMECA] toward Enterprise Architecture to help prioritize which systems to address as well as the urgency for each. I later incorporated it in a revision of the Enterprise Architecture Methodology I co-developed and use today.
Simply put, looking at the Enterprise as a biological entity with various systems that support its health gives you a different perspective on the factors involved with cloud, on-premise, hybrid, etc. Once this perspective is used multiple factors become clear as to the tradeoffs that should be expectedly address.