Ugly Side of Cloud: The Rebuttal
Nearly two weeks, I wrote about the Ugly Side of Cloud Computing. Not long after, vendor responses trickled in.
IBM made a valiant effort to point out that they view multi-clouds as good, and one other asked about the specific definition of a private cloud.
Microsoft’s National Technology Officer, Dr. Dzahar Mansor, also responded to the article, saying,
“I remember the same debates and issues were raised when the world was moving from big machines to client server. Again, nothing is hidden. Everyone needs to do their due diligence and proper risk assessment (business perspective) to decide what is best for them now, and their roadmap to the future. (The) key is to make an informed decision based on deep understanding of their business, their customers/market and how various technologies can help.
“If the Business is not ready for cloud now after this analysis, that’s fine, but do have a plan to ensure we can respond quickly if needed when our competitors start to our-compete us in terms of the cost, agility and security of doing business.
The nice thing is we have a choice.”
An industry source, with a unique perspective of the whole industry, because of his high vantage point and regular dialogues with global cloud players, also responded.
In one fell swoop, all the points made in Ugly Cloud article was rendered moot, because in his opinion, cloud deployments out there today, do not fulfil the 5 cloud criteria of NIST.
The rest of this article is based upon my conversation with him.
First things, first
Things have certainly evolved for the cloud industry, since I first started hearing about it, in 2011. Vendors were continually expounding on the virtues of cloud; that you would only pay for what you used, hence there was no need for capital expenditure, and that the cloud was elastic because you could scale up or scale down resources, and so on and so forth.
Cloud computing (specifically private cloud) was also tied in very closely with virtualisation.
In fact, the biggest virtualisation vendor was big on ‘selling’ the idea, that if virtualisation had phases of adoption maturity, cloud computing and later, the software-defined data centre, was the end goal of every virtualisation initiative by an organisation.
For me at least, the idea about no capex was very viable, because thanks to virtualisation, organisations would be reusing existing hardware. This reduces wastage, cuts down costs, and this means Big Yays all around, right?
Wrong.
This is turning out to be NOT the case.
Back to my conversation with my industry expert.
According to him, if NIST’s cloud computing criteria is to be strictly adhered to, there should be no such thing as a private cloud, because new hardware would have incurred capex, as does software licenses that come with contracts.
One other cloud criteria, dynamic scalability, would also not be possible because it would be limited by hardware; to scale up workloads would require more resources (CPU, memory, etc) and more hardware would have to be put in, which in turn incurs more capex.
This is turn, does not meet another cloud attribute – shared pool of resources.
In these ways explained above,and in this expert’s eyes, only Google, AWS and Microsoft (to some extent) are offering real cloud.
So, where does this leave all the cloud vendors out there, who offer private and hybrid cloud solutions, and their ecosystem of value-adding partners?
A meet half-way solution for now, is that these vendors may not fulfil all five criteria, but at least they may meet part of it.
So, if you can live with that, read on. If you can’t, read on anyway.
Costing of the cloud is a very complex issue. The expert pointed out, “To get a true gauge of the true cost of cloud, you have to take into consideration at least, TCO for up to 5 years.” This is because, hardware depreciation has to be factored in.
To truly gain benefit from cloud, apps also need to be redeveloped, which means more cost. According to my expert source, some businesses do a lift and shift of their environment, from on-premise over to the cloud.
This does not end well for those businesses.
“Re-architecture is needed before you lift and shift workloads from on-premise to the cloud,” he said, while also pointing out that if workloads are flat (as in requiring have a flat demand of computing resources), there would not be much savings that a cloud delivery method, would bring.
To see the cost savings, that I’ve heard so much when I first started reporting about this, workloads need to be the bursty kind. Bursty workloads experience extremely huge peaks at certain times. For example, when online shopping traffic increases after office hours.
Fully Leveraging cloud
The nature of cloud, makes it very, very suitable in certain situations, according to the expert.
“If you are just replicating your backend IT, there is no benefit (to gain),” he said. He also added, that he sees the benefit of cloud, really coming from the toolsets which they offer.
“Clouds offer toolsets to build things bigger; analytics, machine learning, artificial intelligence, blockchain…”
With regards to this, AWS comes to mind for him.
But, despite all the wooing that our government has done to bring global cloud players to our shores, there have been a few setbacks. One however, Alibaba Cloud, has finally landed in Malaysia, and research company IDC expects other global players like AWS and Google and Microsoft to follow suit.
In the meantime, there are private cloud deployments in Malaysia, and there are also businesses here who leverage cloud computing, in one or two or all of the three forms that cloud comes in – platform, infrastructure, and software. There are also local cloud service providers.
On the global cloud player front, Alibaba Cloud in Malaysia, is some positive development, but things are not all peaches and cream, as certain factors, are not allowing the cloud computing industry, to progress as it should, especially so for the regulated industries. This deserves another article.
That said, there is also a lack of pull demand from businesses in Malaysia, and rightly so, because of the mostly negative perception about cloud vendors and cloud deployments that is currently pervading.
My expert. and even other IT heads, have cautioned, “Be careful of vendors with a legacy licensing business.”
This is very good advice, moving forward.
A KPMG cybersecurity lead had observed that there is a glaring lack of good contract managers and legal teams who understand cybersecurity and cyber risks. She had emphasised, “Please make sure you read your contract very carefully.”
This is even more so for cloud computing. Because when you really think about it, the technology is still quite nascent for a majority of businesses in Malaysia, and the challenges posed by regulators, and confusing contracts, don’t help its cause.
RELATED ARTICLES
You must be logged in to post a comment.
There are no comments
Add yours