It is not the actual performance but management of multi-cloud that is worrying both the users and potential users
Cloud has become the primary IT delivery model for today’s enterprise. While different organizations—especially the early movers—may have started with their own reason—CFOs of large organization for turning Capex to Opex, entrepreneurs for competing with larger ones as equals by getting access to the best technologies, and CIOs for simplifying their infrastructure management complexities—today, most seem to have realized that, all these—and more—are valid reasons for imposing their faith on cloud.
When it comes to new ideas, people usually take a path of caution, with a low-risk approach, even while knowing fully well that higher risk may yield a higher reward.
That is pretty common in enterprise IT—especially in large organizations where the downside of a thing going astray is much higher as compared to, say, a start-up. CIOs often go for the ‘low-hanging fruit’ not because they themselves are risk-averse but they know that if things go slightly wrong, even without causing much real damage, that gives room for naysayers to stall things.
Very often, this approach means doing one thing at a time. As such, cloud was a completely new model. It meant, at the very basic, letting out physical control of things. Most started with small experiments with just one public cloud service provider, when it came to their infrastructure services (IaaS). The reasons were many:
It was a new model and the SLA metrics were new. Not many had a visibility to what would happen. Just that the financials were too inviting for people to jump.
It was limited workload and limited set of services that were entrusted to public cloud. So, there was not really a need to look beyond one service provider.
Handling one service provider was considered relatively easier. And since it was just an experiment, it was not worth trying so many permutations and combinations.
Even when they decided to scale up, doing that within the same service provider was much easier than to redo the entire process of learning and discovering another platform.
Last but not the least, where was the choice? For many, it was AWS. And only AWS.
From Cloud-exploring to Cloud-centric
As cloud becomes one of the major delivery models—accounting for anywhere up to 60-70% of the workload for large organizations—the focus has shifted from ease of doing to getting the best out of it.
Enter multi-cloud—which is becoming the de-facto cloud strategy for many organizations. According to IDC’s 2019 Multi-cloud Management Survey, 93.2% of US IT decision makers said that their organizations are currently using more than one infrastructure cloud. Out of these, as much as 81.5% said they were using multiple public clouds while another 7.5% said they were using one public cloud with one or more private clouds.
The reason is not difficult to guess. Especially, the top two reasons in the IDC survey are not surprising. One is a supply side imperative and one is a demand side imperative. As many as 61.5% said they were using multiple clouds because they wanted to leverage cloud-specific capabilities, while 54.5% said the reason behind using multiple clouds for infrastructure is because specific business units wanted so, reflecting the rising trend of users’ choice prevailing, even in tech infrastructure—an important takeaway.
Interestingly, as many (54.5%) pointed to a reduction in cloud spending as a reason for choosing multiple cloud options. To say the least, this is an interesting find and is quite counter-intuitive. Other reasons such as data locality are not really technology decisions, so we can ignore them for the present discussion.
While the IDC survey was restricted to the United States, by far a more mature market, the Rightscale 2019 State of the Cloud report, the latest edition of this annual comprehensive cloud research, that conducts the study among global respondents, also found similar trends.
The Rightscale research found that 84% of the enterprises it surveyed have a multi-cloud strategy, as compared to 81% a year before, while those planning a hybrid cloud strategy grew to 58% from 51% in 2018.
Challenges of Managing Multiple Cloud
IDC research did explore the challenges faced by organizations in managing multi-cloud. But those challenges are specific to the United States—a market that is not just mature in technology usage, it is an efficiency-driven market, as compared to a growth-driven market like India.
For understanding the challenges of multi-cloud management, their reasons and how the organizations are planning to tackle those challenges in India, CIO&Leader conducted a research among 56 senior and middle level IT professionals—mostly CIOs/CTOs/IT Heads and VP/AVP level professionals.
The research was targeted at specific senior level IT executives and was conducted in February-March 2020.
We present here the findings. The research was broadly divided into three parts:
1. Part I was meant to understand the current IT infrastructure of organizations and their usage of cloud
2. Part II, which was the core of the research, tried to understand the challenges faced by those using multi-cloud and envisaged by those planning to use it in future, the reasons behind those challenges as they see it and the possible solutions, as seen by them
3. Part III was a broad outlook for the future: Their cloud management strategy in the medium term – in three years from 2020 to 2022
State of the Cloud in Indian Enterprises
As many as 88% of the organizations have cloud. With a significant number also having on-premise, it seems organizations are moving their managed hosting and co-located infrastructure to public cloud. Only 36% said they have co-located infra.
Hybrid is the default cloud-mix for organizations. As many as 84% of the respondents said they have a proper hybrid cloud infrastructure. More than one-third of them said they have a pretty good mix of public cloud, private cloud and legacy.
More than half of the respondents said they have multiple cloud service providers, with more than one out of three saying their organizations use multiple hyperscale service providers.
However, as expected, for about two-third of the organizations, their workload on public cloud is 40% or less.
So, what are the big challenges in multi-cloud setups? The short answer is: Managing them. It is not the cloud performance that is the big challenge. Lack of seamless portability is not seen as a big challenge; neither is inconsistent application performance. Not, for that matter, is challenge in seamless data flow. In short, multi-cloud, if you ask a simple question, works, without too much of a challenge.
But business is not about making things work somehow. It is about getting the best out of a resource, and the assurance that you are in control. Now, take a look at the top challenges. The biggest three are end-to-end visibility, end-to-end security and complexity of management—all management issues.
It works, but ‘we still do not have the confidence that we can do what we want’. Take, for example, challenge in meeting regulatory compliance. It comes as the fourth biggest challenge, as seen by CIOs.
In short, cloud in cost and performance may be comparable, even better, than legacy, but the control—and that is very important in this risk-laden business environment—is still not adequate.
That is the big message from the survey.
Another interesting aspect from the survey is the difference between the ‘songs of experience’ and ‘songs of innocence’ (with apologies to William Blake). See the difference between the ‘experience’ of those who use multi-cloud and the ‘perception’ of those who have still not started using it.
In some cases, the difference is huge. End-to-end security is perceived to be much better in multi-cloud than what the experience tells the multi-cloud users. Same with cost—the big reason behind cloud getting the buy-in from top management. The users see it as a much bigger challenge than the non-users. Complexity of management, challenge of meeting regulatory compliance, inconsistent application performance—across parameters, and the story is the same: The actual users find them to be bigger challenges than what the non-users perceive it to be.
In general, that is true in case of most challenges. What does it say? Marketing is better than the reality. In short, multi-cloud is still in the hype cycle.
What explains the challenges? Again, management issues come on top. Lack of a single management portal is overwhelmingly seen as the top challenge—by almost three out of four multi-cloud users. No accountability by one service provider and lack of industry standards are the other top reasons. Even lack of inhouse skills are identified as a reason by quite a few. Few think improper design has anything to do with the challenges.
When challenges are many and are primarily to do with management, what can the solution possibly be? Seek the solution in expertise and not in technology. Right? And that is what the respondents say.
A predominant majority prefer both the possible options available: Skilling and outsourcing. The existing users of multi-cloud are a little more pragmatic and hence want to go for external service providers to help them sail through, while those who have not yet got into multi-cloud naturally think they have time to build inhouse skills. But the difference between users’ and non-users’ solution do not vary much. Very few want to do a U-turn, to go back to on-premise. And very few of the existing users of multi-cloud go back to a single cloud provider, though a higher number of non-users see that as a solution to the challenge of multi-cloud.
The big message: People do not want to run away. They want to tackle the challenges head on.
So, going forward, what can we expect, in terms of their cloud management mechanism? Not much surprise here. Almost equal number of organizations wants to do it with the help of external service providers and by using inhouse skills. While 25% believe they have the required skills, 21% say they will go and acquire those skills.
So, how would the cloud set-up in 2022 look like? Hybrid is the way forward, with two-third saying they envisage a balanced hybrid cloud set up.
The Topmost Task
As businesses turn more and more towards technology to not just solve their business problems but also to explore new opportunities, cloud will be the new innovation hotbed. A cautious approach will have huge opportunity cost while things going wrong will have severe downside. So, most organizations will try to get the best out of cloud. And multi-cloud’s TINA factor will only increase.
In fact, CIO&Leader, in its forecast for 2020—Five areas CIOs will spend most of their time on in 2020—had identified Optimizing Cloud Portfolio and Performance, as the topmost area on which CIOs will spend time in 2020. In fact, it will go beyond 2020.
“A variety of reasons—conscious choice, availability of specific applications and mergers & acquisitions—is making organizations settle for multiple clouds,” we had written.
“The CIO’s challenge is to choose the right combination, ensure that they work together coherently for the business to be run on them smoothly and get the desired business benefits—and of course, smoothly keep running it. That is what we call optimizing cloud portfolio and performance,” we had further said.
“Since this is the area that will touch most part of the business and will account for most of the technology investment, we see most of the CIOs’ time will go towards this activity,” we had observed.
Our research had identified this—optimizing cloud portfolio and performance—to be the single-most important activity for large enterprise CIOs in 2020.
We stand by that observation. In fact, the results of this survey, done three months after that article was done, has just vindicated our prediction.
Indians love diversity. And are comfortable managing it. Why should cloud be an exception?
Cloud, in terms of cost and performance, may be comparable, even better, than legacy, but the management control is still not adequate. That is the big message