Cloud First No More, Cloud Smarter Instead
A Request for Comment (RFC) on cloud computing adoption across the federal government has gone out and it says that it’s time to move from Cloud First to Cloud Smart.
According to an article on NextGov.com,
“The new strategy will update the Obama administration’s “Cloud First” policy, established in 2010, to better reflect where agencies and the technology are today. Seven years ago “was a time when cloud computing was still new,” Federal Chief Information Officer Suzette Kent said during a roundtable with reporters Monday. “Many agencies were early in their journey in adopting those technologies and we’ve learned a substantial amount within the federal government, as well as the capabilities in the industry, which have significantly advanced.”
The strategy RFC adds, “Much of the previous guidance on the topic of cloud technology focused on potential benefits instead of realizing outcomes.” And, although consisting of many components, it highlights application (suitability for) migration, security, and workforce. The U.S. government is not alone in its approach. We have been hearing similar approaches from our large customers, who have replaced the lift and shift strategy with refactoring of application code to take advantage of cloud technology and careful planning for security, management, governance all of which depend on good data before, during, and after migration.
In its latest white paper, “A Roadmap for End-to-End Monitoring of Enterprise Applications in Hybrid Multi-Cloud”, ONUG Monitoring & Analytics working group identifies wire data, machine data, and legacy data sources as the basis for enterprise management. The early cloud adopters were at a disadvantage however. Cloud providers only offered machine data; voluminous amounts of cryptic machined data that had to be stored and mined for intelligence after the fact at significant cost.
No wonder that the RFC highlights application (suitability for) migration, security, and workforce! All three were impacted by lack of high-fidelity wire data before and after migration. But there is hope! Ms. Kent is correct in saying, “the capabilities in the industry, which have significantly advanced.”
High fidelity low noise wire data is now available virtually across all cloud providers. This bears a pause and re-examination of earlier tooling strategies. In light of the mandate for smarter cloud adoption let’s examine the impact of high fidelity wire data on each area.
Application (suitability for) migration
This is part understanding the application code for suitability to take advantage of cloud technology and part discovering and documenting all application components, their dependencies, data gravity issues, communication habits of application components and its user communities; and, still part having baselines of user-experience for those user communities. This work has to be maintained after migration on an on-going basis. Yet, outside code review, wire data is the only basis for solutions that provide these answers. If the metric you seek is a byproduct of communication, machine data cannot help you; you need wire data!
Monitoring is part and parcel of any security strategy. Again, without wire data one’s defenses are handicapped. As in the real world, cybercrime travels into your systems, migrates, reaches out for command and control servers, and exfiltrates data. A log entry tells you that your perimeter has been infiltrated; a server compromised. In other words it is telling you that the criminals are already in your house. That is if you happened to have programmed the proper rules to decipher machine data for that particular attack. Conversely, wire data has the ability to catch the cyberattacks in motion and even stop them authoritatively before breaching the premier. Best part: it has a much lower cost of ownership!
A year ago I wrote extensively about the workforce challenges in my blog, “Service Assurance in Hybrid Cloud at an Affordable TCO”. The frequency of research reports sighting skilled staff shortage has only increased recently. One area that you could cloud smarter is through a unified management strategy that extends from your traditional infrastructure into the cloud as opposed to a fractured one. By doing so you leverage your existing workforce with deep domain expertise to manage application services on both legacy and cloud infrastructures. No need for new staff, expensive training and ramp up time, and most importantly retaining highly knowledgeable staff!
Cloud adoption is maturing before our eyes. And, Regardless who’s Cloud Maturity Model (CMM) one considers, the key gates can be summarized into Planning (inclusive of architecture, workforces, cost, governance, management, security, risk & compliance), Implementation (inclusive of monitoring and security), and Optimization (inclusive of cost). Our own ONUG Hybrid Multi-Cloud (HMC) working group published the following cloud adoption framework last year.
While the entire framework is a data-intensive one for decision-making and operations, early cloud had an adverse effect on the availability that data forcing all enterprise to adopt a fragmented tool strategy against a back drop of severe shortage of skilled staffing levels that continue to persist. The issue is further exacerbated by the fact that different cloud providers have different data strategies for monitoring and security data.
According to IDC in 2017, cloud spend by IT edged out spend by Shadow IT or the line of business. This is a good development for the industry. Successful migration to the cloud is of strategic value for to the business. IT is well positioned to step up with all it experience in governance, management and security.