Organisations are always looking for ways to trim cost, either just to take money off the bottom line or to be able to afford to do other, more interesting initiatives that add business value through alignment and agility. If we look back to where the big savings have been made in the delivery of IT services, and in particular in the delivery of IT infrastructure services, we will all recognise these first couple of trends.
First there was virtualisation for consolidation
Boiling it down, companies ran more applications on each physical server they had, and hence saved money on physical hardware. Significant savings were made in hardware investment and datacentre facility operating costs. Some of these savings had to be re-invested in developing new skills to design and manage all the little snowflakes now bouncing around in the virtualisation snow domes; and new third party ecosystems sprung up to manage the hypervisors, security, backup, capacity planning, etc.
Then there was a move into IaaS
The primary motivation here was to save on operational costs. No longer were in-house teams required to maintain the infrastructure layer, nor even the infrastructure services (provisioning, protection, security) in many cases. However, often the anticipated savings were diluted when workloads were moved "as-is"; and so IaaS Right-Sizing followed hot on the heels of the migration in IaaS.
In the IaaS, not As-is blog I mention that there are plenty of other benefits from using higher value service than IaaS, as well. Things like a managed database - AWS RDS and variants - or a managed CRM - Salesforce. So here, I'm using IaaS as representative and a well-recognised term for something delivered as a service i.e. something delivered and managed by someone else so you don't have to!
Now we look to right-source with a multi-cloud strategy
By right-sourcing we mean identifying virtualised server workloads that should be moved to a more appropriate location (provider) based on service quality and cost objectives. There is growing recognition amongst enterprises and government agencies that a multi-cloud strategy is best placed to match the cost, performance, security and service level requirements of whole sets of systems, to the appropriate provider locations.
There is money to be saved without a doubt, as Jim Kelly discussed in a blog at the end of last year. This is what will get the attention and ultimately the support of those wielding the abacuses.
To my mind though, the 'higher' goal of facilitating this right-sourcing is to enable choice. With an expanded arsenal of appropriately vetted and managed cloud delivery models available to IT, application owners, architects and business managers; the most optimal balance of cost, agility and service quality can be designed for any given system or solution. So, as it is in the way of the world, we prove that we can reduce cost, and our compensation is that we get to launch an interesting and rewarding project that will allow us to do what we do best - apply technologies in the best way possible.
Benefits of a multi-cloud strategy
To my mind the benefits of a multi-cloud strategy delivered through a Service Brokering Capability are:
- The right services reach the right users – cloud services that have been ‘discovered’ and used by one business unit can be exposed to others.
- Levels of automation can be increased - moving from templates that deploy an OS, to recipes that provision a system of multiple connected servers with the right versions of applications installed and dependencies configured.
- Lifecycle management can be applied across all IT provisioned systems in the form of leases for transitory components that are only needed for a while, or even periodic spin-down of systems when they are not needed - leveraging automation to effect infrastructure-as-code so sets of systems can be gracefully shutdown and brought up again in the right sequence.
- IT appropriate systems are provisioned so that brittle cloud services (from the point of a rightfully risk averse Board or IT Executive) can be fortified with secure networks, encryption of data, authorisation, authentication and integration to the right core systems.
- Safe experimentation and innovation are enabled and encouraged, with the cost of ‘failure’ diminished.
- And probably most importantly of all, cost visibility is improved so ‘fiscally responsible’ choices can be made – i.e. ensuring that the cost of the service is appropriate to the importance of the service. And with the centralisation of spending, greater buying power can drive down costs.
A capability is needed to support this
When we talk of a Service Brokering Capability, we mean the establishment of a competency where IT becomes the central point of control, provision, self-service and integration of workloads to a variety of suitable destinations (i.e. effecting your multi-cloud strategy). It is important to recognise that the Service Brokering Capability will not be achieved through application of technology alone. Being a capability it requires technology administered by people with the appropriate skills, through aligned processes and governed so as to ensure IT and business appropriate solutions are implemented.
This is illustrated in the graphic below:
For the rest of this blog I'll focus on the technology area of the portal used for the aggregation and provisioning of services from multiple locations. This is referred to as the Cloud Management Platform - a term and category now recognised by Gartner. The growing number of Cloud Management Platforms available on the market provide for the management of public, private and hybrid cloud environments, and allow for the establishment of service chains to form IT services that support actual business needs. An example of this would be the establishment of a business data archiving service, where the curation, policy and security are controlled by on-premises administered middleware, backed by off-premises storage as a service.
Quantify the savings
Being able to establish this capability and the CMP is predicated on the assumption that there are savings to be made - so we need to validate that. Analysis of your IaaS workloads and their service quality needs, contrasted against empirical data points for the cost of consuming various cloud services, and illuminated with modern data visualisation techniques, will reveal where the greatest yields will be returned.
Through our delivery of cloud strategies for a growing number of organisations we have built up an in-depth picture of the New Zealand real world total cost of consumption for off-shore clouds, local sovereign commercial and government clouds, and on-premises delivery models (i.e. what we used to do exclusively before cloud arrived).
If all your workloads are still on-premises (as opposed to IaaS), then the same will likely hold true, but you may need to first overcome some internal barriers to 'cloud consumption' generally, and we'd advocate that the formation of a Cloud Strategy is the best way to achieve that.
Select a cloud management platformTo select a CMP that aligns with your needs, the important attributes to consider will be:
- Where will the CMP reside – on- or off-premises and as a managed service or internally operated?
- Enabling rapid self-service – with service blueprints that may well deploy a service across multiple cloud locations from the outset to ensure service availability.
- Portability across cloud and virtualised environments is supported and enabled.
- Supports DevOps and configuration management.
- Management, visibility and governance across everything from one place for increased visibility of overall IT availability.
- Cost and capacity management – extending ideally to concepts like cost scenario analysis of the impact of moving a set of selected workloads from their present location to a different cloud.
There are savings to be realised by aligning the differing IT service quality needs to the right location (on-premises or a cloud). Understanding and analysis of your workloads and services needs to be done in context of the actual real-world cost for using any of these delivery models.
A Service Brokering Capability is needed that provides the enabling technology (a cloud management platform) administered by people with the appropriate skills, through aligned processes and governed so as to ensure IT and business appropriate solutions are implemented.
The people, process and governance changes won't be insignificant (and we'll talk about these in another post), but these challenges are soluble and begin to dissolve once the savings to be realised are quantified and validated.
Do you have a lot of disparate cloud services across your business? And do you think that if this method proved it could save money, this would be something that would get support from your executives?