The last decade has seen cloud adoption (both single and multi-) soar. Yet according to a recent EMA webinar, only 35% of the 350 IT stakeholders involved classed their multi-cloud adoption as successful. There are a variety of reasons - diminished security monitoring, vendor sprawl, and increasingly complex network designs are but a few. It's easy to see then, particularly for the poor network engineer who has to implement and manage multi-cloud structures, how some are still skeptical about its true value.
For today, we're going to be a little more positive as the autumn gloom sets in by looking at the benefits that contributed to this rise in adoption.
Consider it a trusty field guide for those who might still have their doubts. Where necessary, we'll also consider some of the caveats to the advantages we discuss. After all, nothing is perfect.
Let's first address something that impacts engineers and management teams alike.
Hosting infrastructure on cloud servers means you obviously don't have to spend as much on on-premises hosting - from physical data centers to air conditioning to cool server rooms, et cetera.
Cloud services are based on an Operating Expenses model (OpEx), as opposed to Capital Expenditure (CapEx), which is typically associated with on-premises hosting. OpEx is a type of "pay-as-you-go" model (PAYG), that promotes variable spending depending on how companies use the services they're paying for.
As Jose Moreno said in the Community Fabric episode 'Networking in a Multi-Cloud World', OpEx lets teams create, deploy and delete whenever they want, optimizing how they spend, and making their sizing a more refined model that can be done ad-hoc.
Despite this, cloud-based OpEx models don't always lead to savings. After all, one of EMA's research findings around multi-cloud networking challenges, was that 30% of surveyed organizations struggle with multi-cloud cost management. As an analogy, owning a car (CapEx) may be more expensive initially, with insurance and fuel to pay for, but in the long run, it can be a lot cheaper than taking an Uber (OpEx) everywhere you go.
Depending on the specific vendor(s), and the fact that PAYG models can change their rules or set usage limits, nothing is set in stone, and cost optimization isn't guaranteed. How teams use or plan on using the multi-cloud therefore needs to be planned in detail in advance to maximize returns on the multi-cloud.
Not being anchored to a physical location as with on-premises servers and VMs means that workloads can be distributed further and wider than ever. Whether teams are spread across sites in a single country, or across continents, the cloud improves the efficiency and simplicity of sharing workloads.
Increased mobility thanks to multi-cloud introduces a different approach to automation. Traditional on-premises work often led to teams becoming trapped within silos when attempting to automate network tasks. Within the cloud, teams are able to deploy and spin up infrastructure at the same time in a way more like orchestration than task automation - leading to less "hoop jumping" to try and achieve cohesion across national teams.
Workload mobility means that teams also have full access to services and data across multiple instances of infrastructure. This is particularly helpful during outages. After all, downtime = sunk cost. Ensuring that, when an outage occurs, teams can spin up a backup instance to replicate core services and connectivity ensures data remains protected and available, and means revenue isn't lost.
The downside to this is that increased potential availability also introduces increased network complexity into an ecosystem. Increased complexity which in turn leads to a decrease in network observability. Teams need to find the right balance between increasing global availability and their infrastructural complexity to maximize the value they obtain from cloud investments. (Stay tuned for more on how you can battle cloud-induced complexity with Network Assurance!)
International team cooperation is also simplified in the cloud, as the cloud offers a single API to manage all of their elements (VNets, Kubernetes clusters, etc.) Instead of having to individually configure ports, create VLANs and all of the associated goodness that comes with working across teams, they all have a single access point to work from.
Cloud providers themselves can also chip in here. Providers typically manage routine updates, patches, and maintenance tasks - simplifying matters further by reducing the manual burden on infrastructure teams and sharing risk with them. BUT, this again comes with its own limits. Using multi-cloud introduces the risk of different cloud providers providing inconsistent levels of support - whereas one may go above and beyond to optimize your experience and offer everything you could possibly want to fit your networking needs, another may provide the opposite, leading to part of your multi-cloud being less optimized than it could, or should be.
Perhaps one of the biggest draws of multi-cloud when compared to single-cloud. During the initial cloud adoption 'craze', some enterprises fell into a trap that they didn't see coming. They focused all of their efforts on a single cloud provider. Keeping everything consolidated under one vendor with a single set of compliance regulations to observe seems like a smart move though? Right?
Not exactly. Concentrating on a single vendor ties an enterprise to that vendor. This vendor 'lock-in' reduces application portability across platforms, therefore reducing workload mobility. Whilst a multi-cloud strategy doesn't provide a guarantee for application portability, it definitely gets teams closer than when opting for a single provider that induces dependency and lock-in.
The multi-cloud shouldn't be feared. Whilst it introduces the risk of increased network complexity, and it can become harder to maintain the overall security of a wider network area, it can also greatly increase operational mobility, refine sizing, and ensure that services and functions can consistently be kept online. This (multi-) cloud does indeed have a silver lining.
Stay tuned for a deeper dive into all things cloud over the next few months. If you'd like to find out how IP Fabric can help you maximize how you use your cloud solutions, feel free to reach out for an obligation-free, tailored demo here.
The last decade has seen cloud adoption (both single and multi-) soar. Yet according to a recent EMA webinar, only 35% of the 350 IT stakeholders involved classed their multi-cloud adoption as successful. There are a variety of reasons - diminished security monitoring, vendor sprawl, and increasingly complex network designs are but a few. It's easy to see then, particularly for the poor network engineer who has to implement and manage multi-cloud structures, how some are still skeptical about its true value.
For today, we're going to be a little more positive as the autumn gloom sets in by looking at the benefits that contributed to this rise in adoption.
Consider it a trusty field guide for those who might still have their doubts. Where necessary, we'll also consider some of the caveats to the advantages we discuss. After all, nothing is perfect.
Let's first address something that impacts engineers and management teams alike.
Hosting infrastructure on cloud servers means you obviously don't have to spend as much on on-premises hosting - from physical data centers to air conditioning to cool server rooms, et cetera.
Cloud services are based on an Operating Expenses model (OpEx), as opposed to Capital Expenditure (CapEx), which is typically associated with on-premises hosting. OpEx is a type of "pay-as-you-go" model (PAYG), that promotes variable spending depending on how companies use the services they're paying for.
As Jose Moreno said in the Community Fabric episode 'Networking in a Multi-Cloud World', OpEx lets teams create, deploy and delete whenever they want, optimizing how they spend, and making their sizing a more refined model that can be done ad-hoc.
Despite this, cloud-based OpEx models don't always lead to savings. After all, one of EMA's research findings around multi-cloud networking challenges, was that 30% of surveyed organizations struggle with multi-cloud cost management. As an analogy, owning a car (CapEx) may be more expensive initially, with insurance and fuel to pay for, but in the long run, it can be a lot cheaper than taking an Uber (OpEx) everywhere you go.
Depending on the specific vendor(s), and the fact that PAYG models can change their rules or set usage limits, nothing is set in stone, and cost optimization isn't guaranteed. How teams use or plan on using the multi-cloud therefore needs to be planned in detail in advance to maximize returns on the multi-cloud.
Not being anchored to a physical location as with on-premises servers and VMs means that workloads can be distributed further and wider than ever. Whether teams are spread across sites in a single country, or across continents, the cloud improves the efficiency and simplicity of sharing workloads.
Increased mobility thanks to multi-cloud introduces a different approach to automation. Traditional on-premises work often led to teams becoming trapped within silos when attempting to automate network tasks. Within the cloud, teams are able to deploy and spin up infrastructure at the same time in a way more like orchestration than task automation - leading to less "hoop jumping" to try and achieve cohesion across national teams.
Workload mobility means that teams also have full access to services and data across multiple instances of infrastructure. This is particularly helpful during outages. After all, downtime = sunk cost. Ensuring that, when an outage occurs, teams can spin up a backup instance to replicate core services and connectivity ensures data remains protected and available, and means revenue isn't lost.
The downside to this is that increased potential availability also introduces increased network complexity into an ecosystem. Increased complexity which in turn leads to a decrease in network observability. Teams need to find the right balance between increasing global availability and their infrastructural complexity to maximize the value they obtain from cloud investments. (Stay tuned for more on how you can battle cloud-induced complexity with Network Assurance!)
International team cooperation is also simplified in the cloud, as the cloud offers a single API to manage all of their elements (VNets, Kubernetes clusters, etc.) Instead of having to individually configure ports, create VLANs and all of the associated goodness that comes with working across teams, they all have a single access point to work from.
Cloud providers themselves can also chip in here. Providers typically manage routine updates, patches, and maintenance tasks - simplifying matters further by reducing the manual burden on infrastructure teams and sharing risk with them. BUT, this again comes with its own limits. Using multi-cloud introduces the risk of different cloud providers providing inconsistent levels of support - whereas one may go above and beyond to optimize your experience and offer everything you could possibly want to fit your networking needs, another may provide the opposite, leading to part of your multi-cloud being less optimized than it could, or should be.
Perhaps one of the biggest draws of multi-cloud when compared to single-cloud. During the initial cloud adoption 'craze', some enterprises fell into a trap that they didn't see coming. They focused all of their efforts on a single cloud provider. Keeping everything consolidated under one vendor with a single set of compliance regulations to observe seems like a smart move though? Right?
Not exactly. Concentrating on a single vendor ties an enterprise to that vendor. This vendor 'lock-in' reduces application portability across platforms, therefore reducing workload mobility. Whilst a multi-cloud strategy doesn't provide a guarantee for application portability, it definitely gets teams closer than when opting for a single provider that induces dependency and lock-in.
The multi-cloud shouldn't be feared. Whilst it introduces the risk of increased network complexity, and it can become harder to maintain the overall security of a wider network area, it can also greatly increase operational mobility, refine sizing, and ensure that services and functions can consistently be kept online. This (multi-) cloud does indeed have a silver lining.
Stay tuned for a deeper dive into all things cloud over the next few months. If you'd like to find out how IP Fabric can help you maximize how you use your cloud solutions, feel free to reach out for an obligation-free, tailored demo here.
