Destination Manchester to seek out the views of senior solutions architects, heads of infrastructure, and operations leaders. The topic at hand? The adoption of cloud in all its stripes – public and private, multi and hybrid – with a specific focus on “gaining an advantage with hybrid cloud”.
Held on 27 March at the city’s King Street Townhouse Hotel, this latest event, hosted in partnership with Lenovo and VMware, provided an opportunity to explore the challenges facing effective cloud implementation – and discuss what technology leaders should be looking into next.
Here are five of the key takeaways that emerged from a night of conversation.
1. Cloud doesn’t suit every workload…
But it certainly suits some. As a general rule of thumb, if traffic is peaky and/or data volumes are set to rise exponentially, then cloud is likely to make sense. If, on the other hand, the demands of a workload are unlikely to change significantly over time, then it will almost always prove more cost-effective to stay on-premise. Within those broad guidelines lies nuance, however. For example, two event attendees – one in healthcare, the other in construction – noted that their organisations relied on sharing and processing high resolution images. Excessive transfer costs and storage demands – plus a need for low latency access – means a pure public cloud approach doesn’t make sense. Hybrid might provide a better solution.
2. One way to understand hybrid cloud
Hybrid cloud is often defined in abstract terms. It draws on the best of both public and private worlds, we are told. But what does that mean in reality? Among the most convincing answers offered is hybrid’s ability to allow for a solution – a workload, a service, an application – that sits on premise but is managed by the same tools used for the public cloud services under your control. Familiarity of a single cloud-like control centre drives efficiency and streamlines processes. It also makes effective use of the cloud skillset already inside the organisation.
3. Hyperscalers continue to dominate (but aren’t universally loved)
Three companies overshadow the rest of the public cloud market. Microsoft with Azure, Amazon though AWS, and Google with Google Cloud Platform (GCP) collectively own two-thirds of the market and measure quarterly revenues in the multiple billions of dollars. A straw poll of attendees at our Manchester event indicated that 72% of organisations in attendance use Azure, 43% use AWS, and (unusually) none use GCP. Just under a third of organisations deployed a combination of Azure and AWS. While acknowledging the utility of these platforms, a number of attendees expressed their frustrations with public cloud, citing escalating costs and a lack of agility among their complaints. The latter is an interesting case in point. Flexibility, after all, has long been a key pitch of the cloud vendor – the ability to scale up computing power as you need it, and scale it down when you don’t. The reality is somewhat different, observed one attending tech leader. He said priority compute went to those who paid for higher levels all year round. This left his organisation struggling to respond to seasonal activity. Another said: “Scaling needs engineering and that can bamboozle [even the smartest technologists].”
4. Connectivity is key
One of the biggest risk factors of cloud computing – noted more than one attendee – is its dependency on wide area network (WAN) providers. That is fine if the service providers in question understand the needs of corporate cloud users – high availability, low latency, and consistent throughput. However, that understanding is often lacking, claimed one guest. Why? In part this is a result of skilled technician moving to other parts of the communications industry, including (ironically) cloud. It is also a consequence, the attendee argued, of market consolidation which has resulted in fewer niche players able to deliver bespoke and/or region-specific services. “Connectivity is key to cloud success,” said another.
5. Data sovereignty matters. And not just for personal data
The protection of personally identifiable information (PII) has always posed a challenge for cloud implementation. While computing power can conceivably come from every geography, strict safeguards against the misuse of personal data – stretching back to the 2018 introduction of the European Union’s General Data Protection Regulation (GDPR) – means geographic ringfencing is a must. Public cloud providers are fully aware of the ask and offer sovereign storage and processing services. But it turns out that it’s not just personal data that is sovereign. Certain countries – notably China – regard other information natively generated as belonging to them. With the explosion of internet-connected sensors and edge devices the opportunity to collect in-region data will increase exponentially. And so will conflicts of ownership between nation and organisation.
The next event in this series, Unlocking Growth through the Hybrid Cloud, hosted in partnership with Lenovo and Nutanix, takes place on 24 April, at the South Place Hotel, London.