Multi-cloud: The New Normal
Mr.Carl Eschenbach of Sequoia Capital mentioned in his guest lecture about a Cloud OS that can move workloads around different clouds and also how AI can be used as a unique selling point for cloud vendors to differentiate themselves from their competition. This blog post will firstly give an overview of current cloud technologies then discuss the multi-cloud trend and why AI is necessary in a multi-cloud ecosystem.
Overview of current cloud technologies
As more and more businesses look to use cloud systems to satisfy their IT needs they are faced with the decision to choose between private, public or a hybrid solution. Organizations prefer private cloud when a full scale migration from traditional configurations is required [1]. This is because it does not have the drawbacks that cloud systems are generally associated with, namely data security, flexibility, compliance and regulation. It gives businesses total control over the infrastructure. However, public cloud is a superior alternative for small and medium sized businesses as it is “based on shared physical hardware which is owned and operated by a third-party provider” [2]. This means the infrastructure is owned by many clients and therefore is more cost effective. Public cloud systems can also be more reliable (performance wise) and scalable but is susceptible to regulatory problems if the host server is in another country. A hybrid solution utilises both a private and public cloud customised to the needs of the business. There is however a new trend in building IT solutions which is the multi cloud strategy. It uses multiple cloud services from multiple cloud vendors.
Multi-cloud: An emerging trend
According to a 2016 survey, almost a third of organizations work with four or more cloud vendors [3]. A multi-cloud strategy is increasingly being used by organizations to solve their IT problems. This is not new as another study conducted in 2015 found a majority of IT departments in organizations of all sizes quickly evolving to a multi-cloud strategy [1]. It allows seamless interaction between multiple cloud systems. With multi-cloud systems, organizations can distribute workloads across multiple clouds and with intelligent algorithms it can choose the best possible avenue depending on the task at hand. This solves an ever present issue faced by IT teams as it can deploy workloads using the best suited system. Therefore, it is not uncommon that companies are now utilizing several cloud service providers (CSPs) to improve transparency, flexibility and efficiency of workloads [4]. In addition, it enhances security, stability and reduces latency issues. Additionally, there is reduced bargaining power from the vendor as services are procured from multiple vendors. Furthermore, another benefit of a multi-cloud strategy is that it can easily integrate into existing platforms without significant capital investments allowing businesses to leverage the new system and compete in today’s volatile markets. Multi-cloud systems also facilitate communication between different clouds to reduce risk of failure and set up contingency plans. Ideally, a multi-cloud system requires intelligence and automation embedded into it due to its complex nature.
Integration of AI into Multi-cloud
Multi-cloud deployments are complex as there are several components at play at any given time. As they are becoming a popular solution for organizations, the ability to monitor, run and maintain these complex systems through a centralized unit remains a challenge. Artificial intelligence can solve this problem as there is a substantial amount of data that goes through the system and can uncover information that may affect the quality of service quickly and reliably. This also solves one of AI’s biggest issues by providing a large number of data sets to work with [5]. AI can help customers get better insights of data, provide faster solutions and helps them innovate better. Machine learning algorithms are on a constant iterative loop to automatically optimize the system so they can not only present information but also analyze and find out the root cause of issues and maybe even make rational decisions in the future. However, since data plays such a major role in AI, an excellent data backup and recovery system should be in place to avoid outages.
In conclusion, a multi-cloud strategy is a promising solution to enhance business value, the integration of AI into cloud systems is inevitable and may very well be how cloud vendors differentiate in the future.
References
[1] https://saas.hpe.com/sites/default/files/assets/4aa6-1660enw-new.pdf
[3] https://www.microsoft.com/en-us/download/confirmation.aspx?id=52044
[4] http://www.centurylink.com/business/enterprise/blog/thinkgig/evolution-cloud-workload-management/
Cover Image: https://www.group50.com/wp-content/uploads/2017/06/Multi-Cloud-Computing.jpg
2 comments on “Multi-cloud: The New Normal”
Comments are closed.
marketing and value chain management. I like to keep things simple and minimalistic!
I enjoyed your post on how companies look to implement a multi-cloud strategy as they navigate the world of cloud computing and how AI could help. It’s an interesting phenomena that among a company’s initial thoughts in the cloud world was to add a layer of abstraction between them and the cloud provider. I’ve always wondered why. Was it to simply hedge their bet on which cloud provider (AWS, Azure, IBM, Google, etc.) would still be around in the next few years? There have been providers that have exited the public cloud space – Verizon (http://ow.ly/I4W930dFVOE), AT&T (http://ow.ly/BTms30dFVPL), VMWare (http://ow.ly/tfBB30dFVQW), Cisco (http://ow.ly/xRdL30dFVRt), HPE (http://ow.ly/i7Wv30dFVSA), etc. So maybe these companies are being cautious.
Another reason that companies may be utilizing a multi-cloud strategy is to retain their control over the infrastructure. Prior to public cloud all the company’s applications ran on servers that were owned and operated by the company. While it was the application that enabled productivity, it was the server that was king. Different business departments had to build to what infrastructure was available or could be procured with whatever CAPEX was available. For that reason, the IT department was run with an infrastructure first mindset. Now with the advent of cloud computing some companies are still thinking the old way. It would be great if the company could abstract infrastructure from a plethora of sources and utilize a multi-cloud broker model where the infrastructure is a pure commodity. Jim Battenberg, CenturyLink discuss this promise on Forbes.com and why this isn’t a reality (http://ow.ly/7xRD30dFVNC). In short, this hasn’t panned out for two reasons: it is very difficult to make that work, and security of the application and the app’s data is a big deal.
But I think this begs a larger question – do companies need a multi-cloud strategy? Or should they pick one main cloud provider. Second, should the application developers have a larger role in IT with the advent of cloud computing; especially considering PaaS services?
Users who have LIKED this comment:
Well written article. Kudos on that.
Cloud services in recent times has caught steam and is on course in becoming the next big thing.
I’m saying on course, even though the industry is worth billions because it is still growing at a rapid pace. Amazon predicts the business to grow at 30 – 40 % YoY for the 3 – 4 years.
A decade ago no one would have thought that this industry would ever come into existence but today here it is and also expanding its base from just providing storage to now AIaaS.
I’m not a technical person and hence cannot speak on the technological aspect or side of it.
Having a business background I can say that the business will continue to grow and economies of scale will rise making it further more viable and also cheaper for the consumers of this service.
Shristi Modi, student at MS&E 238A