SaaS costs and business model optimization to master the digital transformation
digital transformation
SaaS business models, cloud costs and digital transformation are our strengths
biz.space
From the idea to the Software as a Service business model to the costs
tech.talk
Infrastructure-as-Code technology solutions and tools for an optimized SaaS architecture
digital transformation
Thank you for your interest in m2sphere project. We are specialists when it comes to the architecture of native cloud services. Therefore our focus is on Software-as-a-Service business model development, cost and scaling forecasts as well as automated infrastructure as code.
Digital transformation, Software-as-a-Service (SaaS), business models, digital twins, and cloud computing are on everyone’s lips. However, the pitfalls along the way, which greatly impact subsequent operations, are rarely discussed.
This is precisely where we come in. We are experts in cloud cost management by design to ensuce SaaS business model that works!
A major issue that we know from our many years of experience is that in the decision and design phase of cloud services, the right questions are often asked, but as a rule they are only answered superficially, insufficiently or incorrectly.
:: the digital transformation as example
A manufacturer of IoT devices wants to display the online status of the devices in a mobile app. Initially, without thinking twice, he decided on the simplest way to send the status to the cloud every second, as the amount of data is very small. Consequently, this means that a device sends 1,440 messages every day. The manufacturer sells 50,000 of these devices per year, which means 72,000,000 messages per day. As a result, in the first year alone, the manufacturer has to spend tens of thousands of dollars on this function. Therefore, in order to get the cloud cost management under control, a refactoring of this cloud service will be unavoidable. Ultimately, this example shows how quickly a SaaS business model can fail.
:: our USP - SaaS business modeling
We are working on cloud optimization tools that identify cost and scalability issues early in the Software-as-a-Service business modeling process to ensure a smooth go-to-market. Therefore, we use a model of the solution to predict and optimize the operating costs of the cloud solution. Also, scalability is analyzed so that the results can be directly incorporated into business modeling before a single line of code has been written. However, the USP is even greater. The model is used to automatically generate infrastructure code for provisioning and configuring the required cloud components and services. Otherwise, this has to be done and maintained manually, which is a huge amount of work.
:: customer segments
With this new and innovative approach, we are targeting all companies that are considering a transformation of their business model and are planning to host a Software-as-a-Service solution on platforms such as AWS, Azure or Google Cloud. Thanks to sophisticated reverse engineering, companies that already have a solution are also an important and attractive customer segment.
:: our motivation
Get in touch with us
Click HereStill not convinced?
SaaS business model - risk factors
a) Lack of or limited platform knowledge of the cloud providers in question.
b) Cloud cost management and the determination of operating costs cannot be analyzed in advance. The fatal consequence is that cost assumptions are made on the basis of the cloud providers’ static calculations.
c) Solution architects and developers are hardly involved in this phase because it is all about the business.
d) The business model exists as an idea, but the details that matter have not yet been worked out. It is often an internal marketing paper.
What happens if the cloud business model fails?
If one or even several of the above points apply, then the nightmare begins in the first few months of operation, which then show that the business model is not sustainable and time-consuming and cost-intensive refactorings follow. In addition we see this problem rolling towards the industry, whose business can be endangered by unsuitable strategies for digital transformation.
For this reason we have prepared a case study as digital transformation example for you. It shows an easy-to-understand scenario and illustrates the rapidly increasing complexity. This illustrates the problem and shows that the cloud resources used have different billing models and pricing tiers that are not so easy to calculate.
Addendum
If you now say, that’s why we have an external software service provider to take care of this, then we reply: “They will never start their collaboration with you by pointing out the problems first!” After all, software service providers are judged by the customer by how quickly they implement features that contradict a cost-optimized, future-oriented design approach!
:: our experience
In our professional experience, the decision to set up a cloud service is often made with little knowledge of the factors already described. Consequently, the nightmare begins with operation, which often shows that the SaaS business model is not sustainable. As a result, time-consuming and cost-intensive refactorings follow.
- Without sound platform know-how, no valid operating cost models can be created in advance.
- Also if the architecture of the microservices is still completely unknown.
:: our mission
We want to help you extract your business know-how and derive an abstract model of cloud architecture that also takes into account cost and key proposition aspects for a successful digital transformation. Our experience will help you to minimize risks and develop a sustainable business model.
Our services and cloud optimization tools are designed to help decision-makers, solution architects and software developers. In addition to modeling the solution, the added value lies in the automated provision of the multi-cloud infrastructure.
:: our solution
As part of the business modeling, we outline an abstract model of the microservices without needing specific platform knowledge. Additionally, using iClients from m2sphere, we stimulate the system in a user-oriented manner and carry out a simulation based on the target platform. Consequently, we provide facts answering the following questions:
- How high are the operating and cloud costs?
- Does the solution scale as desired?
- How do the platform resources need to be configured?
:: help us
At the moment we only have the idea, we are still private individuals and there is no professional company behind the project yet. We have clarified the feasibility, even if it is technically demanding, through our own tests and in many discussions with relevant university chairs and research institutions.
However, we still need market figures to confirm that companies recognise the problem and are willing to spend money on such solutions. We also need investors, because we can’t finance this on our own.
:: your experience
We believe that the issues we have been working on and trying to solve for some time are a headache for many companies. This is based on our own experience over many years, but it is also the result of a large number of conversations we have had. Therefore, we are trying to find out if this is really representative by publicising our project.
You are invited – indeed, we ask you – to let us know if you have had similar experiences. We are happy to help you with advice and would publish your story and possible solutions – without revealing your identity – in a blog.
:: infrastructure
An essential part of the concept is that the early model is more than just a drawing. Moreover, this sets our approach apart from most tools. Code can be generated from the platform-independent schema right from the start, enabling the automated deployment of the necessary cloud infrastructure.
Furthermore, developers extend and customize this in a scripting language with modern design patterns. This enables the provision of resources and services in multi-cloud environments. Through automatic PaaS translation for AWS, Google Cloud, and Azure, the process becomes seamless.
:: biz.space
Our web presence addresses different tasks and roles in a company, which we believe are inextricably linked togehter to successfully master a business digitalization.
In this section, we focus on the aspects that are directly related to business modeling. Specifically, we show best and worst practices, give tips that should at least be considered, and try to help with advice on the path to corporate digitization.
:: blogs
We created the website to get our ideas out to people and to get the opinion of the potential market. To get something, you also have to be prepared to give something.
Therefore, why we’ve set up blog pages where we continuously share our experience and knowledge on various topics related to digital transformation, business modeling, managing cloud costs and infrastructure as code.
:: tech.talk
In addition to business modeling, future-oriented implementation also plays a key role and can also determine the success or failure of a transformation strategy.
For this reason, a separate section is dedicated to the development-relevant details of the cloud infrastructure, its automated provision and the best and worst practices when using cloud resources and services.
A central heating system reports a serious fault to the cloud service. This error message leads to the owner’s contact details being read from an SQL database, who is then informed of the problem by email and push notification. The error data is stored in a memory. In addition, this data is transferred anonymously to another memory, which is cyclically loaded into a data warehouse in order to be able to evaluate availability figures from operation via Power BI.
In the meantime, the person notified has read the push notification and opens the mobile app to obtain further information. Due to the severity of the problem, the owner uses the app to authorize the heating manufacturer’s customer service team to access the system remotely in order to rectify the fault, thereby triggering a support ticket. A support employee from the company logs into the system and starts a self-diagnosis run. Whereupon …
-
Example
-
Event Hub
-
Functions
-
Storage
-
SQL DB
Digital transformation example
Azure Event Hub pricing
Tiers in detail
| Basic | Standard | Premium | Dedicated* | |
|---|---|---|---|---|
Capacity | $0.015/hour per Throughput Unit*** | $0.03/hour per Throughput Unit*** | $1.233/hour per Processing Unit (PU) | $6.849/hour per Capacity Unit (CU) |
Ingress events | $0.028 per million events | $0.028 per million events | Included | Included |
Capture | $73/month per Throughput Unit*** | Included | Included | |
Max. Retention | 1 day | 7 days | 90 days | 90 days |
Storage Retention | 84 GB | 84 GB | 1 TB per PU | 1 TB per CU |
*To understand how much a CU can achieve, please refer to our benchmark results. Usage will be charged in one-hour increments with a minimum charge for four hours of usage. For more info on this offering please read Event Hubs Dedicated Overview. For additional questions please contact Quota Increase Support or please contact the event hubs team.
**Message retention above the included storage quotas will result in overage charges.
***Throughput Unit provides 1 MB/s ingress and 2 MB/s egress.
Azure Fuctions pricing
There are several ways in Azure to run functions. Two variants are considered here, the consumption-based and the premium plan.
Consumption-based: The costs are billed according to GB seconds used, i.e. the RAM memory requirement (rounded up to 128 MB blocks) multiplied by the runtime (min. 100ms). The number of calls will also be considered.
Premium plan: Billing is based on the number of core seconds and the RAM allocated to the instances. There are no charges per execution. At least one instance must always be allocated per plan.
The difficulty here is that at the design stage there is no data on how long the runtimes are, nor what memory requirements are necessary. Therefore it isn’t possible to determine which variant is the right one.
Tiers in detail
| Consumption-based meter | Free Grant (per month) | Pay as you go |
|---|---|---|
Execution Time | 400,000 GB-sec. | $0.000016 per GB-sec. |
Total Executions | 1 million executions | $ 0.20 per million executions |
| Premium plan meter | Pay as you go | 1 year savings plan |
|---|---|---|
vCPU duration (ref. SKU) | vCPU: $ 0.169 vCPU/hour | vCPU: $ 0.14027 vCPU/hour |
Memory duration (ref. SKU) | Memory: $ 0.0121 GB/hour | Memory: $ 0.010043 GB/hour |
| SKU | Cores | Memory | Storage |
|---|---|---|---|
EP1 | 1 | 3.5 GB | 250 GB |
EP2 | 2 | 7 GB | 250 GB |
EP3 | 4 | 14 GB | 250 GB |
Azure Blob Storage pricing
Azure Blob Storage offers different price levels depending on the access tier (Premium, Hot, Cool, etc.) and redundancy option (locally redundant, zone redundant, geo-redundant, etc.). The hot access tier is optimized for frequently accessed small amounts of data and the cool tier for less frequent access with larger data.
The data size and can still be estimated relatively well for many applications. The number of accesses, on the other hand, is much more difficult and depends heavily on the design
It should be noted that the settings apply to the entire storage account, i.e. it must be carefully considered which data should be stored with which redundancy.
Tiers in detail
The following overview refers to locally redundant storage (LRS). Prices increase with higher redundancy settings:
Zone Redundant Storage (ZRS), Georedundant Storage (GRS), Geozone Redundant Storage (GZRS).
| Operations | Premium | Hot | Cool | Cold |
|---|---|---|---|---|
Write operations (per 10,000) | $0.0228 | $0.065 | $0.13 | $0.234 |
Read operations (per 10,000) | $0.0019 | $0.005 | $0.013 | $0.13 |
Iterative Read Operations (per 10,000) | $0.0228 | all three tiers: $0.0052 | ||
Iterative Write Operations (100’s) | not applicable | all three tiers: $0.065 | ||
All other Operations (per 10,000), delete is free | $0.0019 | $0.005 | $0.005 | $0.0052 |
| Data storage (pay as you go) | Premium | Hot | Cool | Cold |
|---|---|---|---|---|
First 50 terabyte (TB) / month | $0.15 per GB | $0.018 per GB | $0.01 per GB | $0.0036 per GB |
Next 450 TB / month | $0.15 per GB | $0.0173 per GB | $0.01 per GB | $0.0036 per GB |
Over 500 TB / month | $0.15 per GB | $0.0166 per GB | $0.01 per GB | $0.0036 per GB |
Azure SQL database pricing
Determining the purchasing model for databases and the associated scalability is one of the most difficult tasks in the design process. Regardless of whether the vCore or DTU model is chosen, questions and uncertainties usually remain.
Behind the vCore is a calculation based on the computing time consumed per virtual core per second. The computing time depends on the hardware used and the query complexity and therefore leads to the costs.
With the DTU model, the costs are fixed, but it is unclear what DTU performance is required. A DTU (Database Transaction Unit) is an aggregate unit of measure in Azure SQL Database that combines CPU power, memory and I/O into a single metric to measure and scale database performance and capacity.
Tiers in detail
There are basically two different billing models for SQL databases: virtual cores (vCore) and database transaction units (DTU). Further corresponding setting options can be found in the first table.
| Purchase Model | Computer Tier | Service Tier | Hardware Type |
|---|---|---|---|
DTU | not applicable | - Basic - Standard - Premium | not applicable |
vCore | - Serverless - Provisioned | - General purpose - Hyperscale | - Standard series - DC series - Premium series - Premium series, mem-optimized |
Since there are so many advanced, price-influencing settings in the vCore tier, we decided to list the DTU prices (very common).
| DTU Model | DTUs | Inluded storage | Max. Storage | Pricing |
|---|---|---|---|---|
Basic | 5 | 2 GB | 2 GB | $0.0068/hour |
Standard S0 | 10 | 250 GB | 250 GB | $0.0202/hour |
Standard S1..S11* | ||||
Standard S12 | 3,000 | 250 GB | 1 TB | $6.0488/hour |
Premium P1 | 125 | 500 GB | 1 TB | $0.6251/hour |
Premium P2-P14* | ||||
Premium P15 | 4,000 | 4 TB | 4 TB | $21.5054/hour |
* refer to source
Cloud pricing conclusion
We hope that this example has helped you to better understand the underlying problem. Of course, we are aware that the price tables are anything but easy to read, especially if you have never dealt with these topics before. But that is precisely our aim, to illustrate the problems and the complexity. While there are tools such as the Azure Pricing Calculator, they cannot be used to calculate dynamic relationships.
Visit out digital business transformation blog!
We are constantly writing new posts and interesting articles on topics such as cloud infrastructure and cost optimization, SaaS business models, Azure and PaaS pricing and cloud cost management in general. Specifically, we address developers, solution architects and decision-makers in digitalization projects. Furthermore, we would like to provide you with examples of the digital transformation of Software as a Service to further strengthen your success story.
Thank you for your interest!
