Should the cloud now be renamed as something else? Larry Ellison famously once called it "water vapor". 'Air' perhaps, something that everybody uses and nobody thinks about on a day-to-day basis?
This was the one question Nutanix CTO, Sunil Potti, did not pose at the recent London end of the globally-rolling Nutanix.NEXT conference, though he came a good deal closer than ever before.
He suggested that terms such as hybrid cloud were now starting to miss the point, not least because what users are starting to look for is on-demand cloud services where 'the service' was made up of whatever applications and tools were needed for the task:
For example, one goal is to make AWS invisible within a cloud service mix.
Indeed, the goal has to be to make every individual element of a cloud service 'invisible'. It is of course important that each element is there, but increasingly for users it has to be de-cerebralised – unseen and unthought about.
There are a couple of reasons for this that he alluded to. One is that there are constraints to the growth of large public data center – that there are physical laws in play.
One factor here is that the larger the data center the larger, and more geographically-dispersed, the marketplace of consuming businesses and consumers becomes. This has the potential to introduce all kinds of problems such as latency errors and holdups, network bottlenecks and power supply breakdowns.
This must also be set against the second reason Many business users are starting to realise that they need much more dispersed computing resources as they appreciate the need to have processing closer to the edge of their networks – out there at the coalface of their working world.
Having that dependent on a large centralised mega-datacenter has the real potential of taking business risk management back to the earliest days of a single mainframe – and even then it was a while before they were used for mainstream business management for that reason.
Wake up, MSPs
A wider range of service providers will therefore be necessary, which Potti sees as a good opportunity for the existing Managed Service Provider (MSP) community. He does, however, see some issues with their ability to rise to the challenge that now exists:
Some of the early cohort of MSPs may have got it, but many have long term contracts with customers, so their interest is primarily in making more money from those same customers. They are going with Nutanix to save costs while delivering the same service. They are also using it to cut staff. We are seeing more of that.
But we are not really seeing the MSPs develop new stacks in order to provide new services to new customers. That is the opportunity for them now: services such as Desktop as a Service, Disaster Recovery as a Service are obvious candidates.
Nutanix has also developed into quite an acquisitive beast, and as one the primary judges of what to buy/not buy, Potti’s thoughts on the subject, and particularly where the line on acquisitions gets drawn, are important:
We are going to wing it for a while. We are looking for technologies and expertise rather than revenue streams when it comes to acquisitions. Also, it is whether the product/technology in question stands alone and whether it is used by many existing customers? Then we will just be a customer.
It also depends on our customers’ perceptions. Take firewalls for example. Users expect 2-way management of their services to be done within the infrastructure, and that is where acquisition can make sense. But the defence of the data center is best handled by external, third party services.
There is also the potential, if it’s place in the market develops as it would like to see, for Nutanix to be seen as the 'sole arbiter' of what applications and services get to work in the cloud:
We need to see what happens with the possible role of cloud broker in the marketplace. But we do see it as a possibility and not necessarily an unintended consequence. We are aware of the possibilities. But there will be greater dispersal of cloud services, such as MSPs, going all the way out to the edge. Our assumption is that the world will be multi-cloud.
New tools for better glue
The pace of development at Nutanix means that there is nearly always something new to talk about and demonstrate at each emergence of .NEXT, and London was no exception. There were some new tools to be demoed and some now-more-refined implementations of previously announced ones. The underlying goal with all of these is to improve and extend the availability of the `glue’ needed to assemble services of increasingly greater richness.
XRay – a performance evaluation tool – covers both Nutanix and competitive equipment. This makes performance and operational comparisons of applications possible, and represents a public acknowledgement of the obvious - that Nutanix is not always the best solution of any particular workload.
There are also tools for migrating existing apps to the new environment. This, in essence, fully automates the whole process, including the orderly final shutdown of a legacy VM in which the application is running. Migration is said to require just a few clicks to choose such parameters as the type/time of migration and to name the migration process.
Smart resource management systems have been introduced that use AI and machine learning to add a degree of predictive provisioning within Prism. The goal here is to manage inefficient over- and under-provisioning of resources such as VMs. Prism now also has micro-segmentation tools that add greater network visibility, allowing the provisioning of specific network services such as firewalls. These can also be used to discover what devices or services are on the network, the relationship of one to another, and the communications that are passing between them.
As the technology of cloud gets subsumed beneath the business service requirement of ‘on-demand ubiquity’, the availability and sources of that ubiquity start to take on an increasingly important role. And that role presents an increasing opportunity for both the MSPs and the established Value Added Resellers.
The fact that Potti sees MSPs being far more focused on using hyper-converged technologies to just cut costs on existing service provision contracts is profoundly sad – both for the MSPs themselves and for their customers, who perhaps should be looking at their MSP partners for some leadership.
As the technology and tools 'glue' Nutanix is offering continues to get richer the ease with which each side of that equation can engineer their transformation solutions will increase. But it should be happening now.