I love cloud computing. I love the economics and the agility of the cloud and how it changed IT forever. The cloud has solved some of the headaches of IT, notably the silos in operations, the silos in development and the silos in infrastructure.
The virtualization and abstraction of rigid infrastructures and on-premise operations have given birth to X-as-a-Service and Cloud Services. Along with this, comes cloud orchestration, cloud automation, policies, DevOps and plenty more. IT responds well to this and thus, public clouds services like Amazon Web Services, Microsoft Azure, and Google Cloud Platforms are dominating the landscape. Other cloud vendors like Rackspace, SoftLayer, Alibaba Cloud are following the leaders pack offering public, private, hybrid and specialized services as well.
In this pile, we can now see the certain “camps” emerging. Many love Azure Stack and many adore AWS Lambda. Google just had their summit here in Malaysia yesterday, appealing to a green field and looking for new adopters. What we are seeing is we have customers and end users adopting various public cloud services providers, their services, their ecosystem, their tools, their libraries and so on. We also know that many customers and end users having several applications on AWS, and some on Azure and perhaps looking for better deals with another cloud vendor. Multi-cloud is becoming flavour of the season, and that word keeps appearing in presentations and conversations.
Yes, multi-cloud is a good thing. Customers and end users would love it because they can get the most bang for their buck, if only … it wasn’t so complicated. There aren’t many “multi-cloud” platforms out there yet.
On the storage or data management “infrastructure” end, NetApp Data Fabric touts to be multi-cloud but it isn’t there yet. They are still working on the “glue” with SnapMirror and others techs to link the data islands together. SNIA touts CDMI (Cloud Data Management Interface) but it takes time to be adopted and mature. I have also seen the emerging prowess of Elastifile Cloud File System at Storage Field Day 12 in March this year and still get some updates on Hedvig on their Universal Data Plane.
Meanwhile, I am not a good follower of configuration management and orchestration tools such as Ansible, Kubernetes, Chef, Puppet and more. My knowledge is limited and my experience and skill even less. I am aware they are becoming very prominent in attacking the problem of getting the VMs, the containers, the microservices, the DevOps to be put together under one roof, so to say.
All these development and technology are great, but can they be adopted and mature fast enough to the demands of multi-cloud?
As this debate rolls on, the notion of IT being siloed and locked-in (again!) to a specific cloud service provider is becoming a real issue. The lack of interoperability and a common framework for data management and data services are real, and so far, I have seen technology vendors jostling and jousting to be the standard of multi-cloud. I may be entirely wrong here, but we are also seeing differing camps supporting one cloud technology over another. This is much to the dismay of the customers and end users, having to take a side to the cloud services and not able to realize the potential of multi-clouds.
In the end, at this present moment, the cloud itself is causing IT silos, when in the first place, it was to reduce or eliminate IT silos on-premise. For now, the silos are happening in the clouds.