The cloud is increasingly present in companies.
According to data from the Cloud Transformation Observatory, cloud adoption is growing steadily (+16% compared to 2020) and is not limited to SaaS services - which have driven the market during the health emergency months - but also PaaS and IaaS, especially services that relate to application modernisation, orchestration and management of information systems.
Organisations that decide to migrate their systems to the cloud do so with the aim of improving the management of their IT infrastructure. In this way, they can dynamically access virtual resources that are potentially unlimited and free from maintenance costs, as well as delivered according to agreed service and performance levels.
A cloud migration project involves calculating the amount of resources needed to run the applications one wishes to migrate. Yet, most companies do not have a targeted strategy.
Usually the choice of cloud plans is based on the characteristics of the existing hardware or the data provided by the virtualisation environment, but this rough estimate ends up including unneeded or incorrectly mapped resources (unused storage, switched-off virtual machines, etc.) in the migration assessments.
Added to this is the habit of companies to request more resources than budgeted, so as not to run the risk of having to make extra budget requests in the short term.
The result is an overrun that coincides with a considerable financial outlay. Gartner has mapped this waste: according to the research "Top Ten Moves to Lower Your AWS Costs", companies that do not size correctly end up spending 70% more during the first 18 months.
In order to eliminate costs related to wasted resources, a Cloud Assessment is needed, i.e. an in-depth analysis of infrastructure consumption. This assessment is necessary in order to calculate the number of resources actually needed for the cloud migration, and has no small benefits. It makes it possible to:
In order to size correctly, the evaluation of the resource consumption of the infrastructure should be done not on an indicative amount of resources calculated on the capacities of hardware and local server machines, but on the actual daily utilisation of the infrastructure.
How to calculate this? How to produce useful considerations for carrying out the migration as efficiently and economically as possible?
We provide our customers with a Cloud Assessment that takes into account five steps:
With the support of specific software that has an agentless scanning method, we analyse the consumption of the infrastructure for a pilot monitoring period of at least 30 days. This duration is normally necessary to define most PBAs (Pattern of Business Activity), which are processing load profiles linked to the demand for IT services by business activities.
They are the main tool Demand Management has at its disposal to estimate the computing power required to deliver services and are therefore indispensable to anticipate and/or influence demand.
We then collect data and information on hardware, software, network traffic and actual consumption and consolidate them into the Consumption Analysis Report. This document calculates the TCO (Total Cost of Ownership) of the analysed infrastructure in all timeframes and takes into account multiple factors, such as the relocation of possible worldwide locations and application interactions.
This is the starting point for a more in-depth analysis, the Cloud Readiness&Profile Sizing, which maps in detail which services and systems can be migrated to the cloud and the recommended sizing based on consumption.
Other elements are taken into account here, including the compatibility of services with the cloud, disaster recovery capacities (disk space, bandwidth utilisation, site in different regions than the current one, etc.), backups (based on current disk occupancy, different retention and frequency scenarios are assumed), networking (outbound values from cloud systems to the outside world), the economic impact of public IPs, etc.
The information obtained allows us to make a resource sizing based on CPU, RAM, IOPs and network throughput 99% accurate. At this point, we can choose the most suitable cloud plans in a strictly scientific manner.
Cost comparisons are made based on the choice options (on-demand or with annual or three-year pre-allocation) of the different cloud providers and based on the expressed business objectives. Hybrid and multi-cloud environments can also be envisaged, with services and applications running locally and on multiple public cloud providers, in order to exploit the costs and benefits of each platform.
In the course of our analysis, we also map the network-level relationships between the analysed systems. This allows us to indicate to our customers which units can be migrated in bulk. This approach - known as “smooth migration” - which enables part migration of systems, is especially suitable for the implementation of hybrid cloud solutions in lift & shift mode.
The accuracy of the Application Mapping makes it possible to migrate part of the infrastructure without any disruptions or additional costs.
The information obtained also allows us to check which systems and applications are close to EOS (End-of-Support). This is an issue that is often underestimated by companies, but which could have serious operational impacts in the event of external attacks or internal failures.
Migration to the cloud could therefore be the right opportunity to overhaul all critical systems and equip themselves with better performing and more secure infrastructures. It is therefore a technological improvement that generates important benefits in the short, medium and long term.
Costs, priorities, tasks to be done... the customer receives everything he needs to plan the infrastructure migration correctly.
In addition, the cloud has such characteristics that require ad hoc financial management and we also offer FinOps consulting to map the consumption of the new cloud infrastructure and allocate resources to the various cost centres.
A public sector company approached us because it wanted to migrate its Disaster Recovery system consisting of 68 servers, 61 of which were Windows-based and another 7 Linux-based, to the Cloud.
The initial sizing, based on the nominal characteristics of the Virtual Machines, required resources at a monthly cost of more than 30,000 euros. An expense that could certainly be optimised with a Cloud Assessment. We therefore monitored consumption for a period of 30 days and identified several opportunities for savings:
Our analysis identified several systems that could be reduced in terms of computational resources and peak usage: in this specific case, they could be covered by the characteristics of Azure B-Series profiles. This made it possible to lower the incidence of the computational part from an average hourly cost of almost EUR 0.40 to just over EUR 0.22.
By measuring disk data and IOPS, the storage was correctly sized and largely converted from premium (SSD) to normal (non-SSD), guaranteeing performance requirements with a cost reduction of more than 40 per cent.
The reduction, in some cases, of the number of cores also reduced the impact of some software licences (e.g. Microsoft SQL Server) that base their purchasing model on this hardware feature.
The graph below shows the detail of the computational optimisation. Given the dimensions shown, the greater the concentration in the upper right-hand corner, the higher the costs will be; conversely, the more the line is positioned in the lower left-hand corner of the graph, the lower the costs will be. The red line illustrates the analysis based on nominal hardware sizing, while the blue line shows the situation at the end of measurement and optimisation.
As a result of all the changes recommended within the Cloud Assessment, the costs for our client's Disaster Recovery system were reduced by more than 43%, a figure that, considering the three-year contract term, equates to a saving of EUR 417,000.
Insights
OUR OFFICES
OUR OFFICES
PADUA
Via Arnaldo Fusinato 42, 35137
MILAN
Viale Enrico Forlanini 23, 20134
ROME
Viale Giorgio Ribotta 11, 00144
Copyright © 2022 WEGG S.r.l. • P.I 03447430285 • C.F. 02371140233 • REA 311023
Certified company ISO 9001:2015
