Home / Products / MIRRA™



MIRRA™ Real-time Task Scheduling

Many reports predict that in the near future, a countless number of IoT devices will be used for different purposes. Moreover, most of them are real-time applications sampling statuses periodically, sending data to remote application servers. Since the number of deployed devices will be huge, and data processing demands may be massive, using a public, private, or hybrid cloud is not a viable option for ideal server deployment. In addition, managing and optimizing virtual machine infrastructure size is a serious matter to significantly reduce IT costs and increase productivity.

Amerra has designed and developed real-time task scheduling algorithms as part of a middleware (called MIRRA™) to resolve these issues in the form of a smart and dynamic virtual resource controller. MIRRA can operate inside virtual machines within most cloud infrastructures and independently from your application server running in the same virtual machine by providing the real-time task scheduler. Multiple peer-reviewed research papers about the technology and successful tests against existing cloud infrastructure have proven our approach.

For additional information on the technology, please read our FAQ section & our most Recent Publications.

The MIRRA™ Pilot Program

Engineers, entrepreneurs, and pioneers,

Amerra is seeking IoT companies within any industry to collaborate on pilot projects involving the development and release of a generic real-time deployment platform. As you know, many reports predict that in the near future, a countless number of IoT devices will be used for different purposes. Moreover, most of them will be real-time applications sampling statuses periodically and sending data to remote application servers. With the number of deployed devices increasing exponentially, the massive data sizes  and coordination of types of data using public, private, or hybrid clouds will not be true viable options. Why? Due to a serious issue of managing & optimizing their virtual machine infrastructure size that now must be addressed in order to reduce heavy IT costs and increase productivity & profitability!

The implementation of the MIRRA™ commercial version from our research will enable additional external validation by deploying your device and/or program using our middleware to improve your company’s bottom line and IT resource allocation.

This opportunity is zero-risk, as Amerra will implement your application servers into a cloud infrastructure after assessing your business logic and current setup, as well as future potential needs . We invite all interested parties to please fill out the form below, or call 713-529-9776.

We look forward to empowering your company with our revolutionary cloud resource-optimization technology!

The Amerra Team

MIRRA™ Pilot Inquiry

Please fill out the following form
& a PDF of the Pilot Summary Overview will be in your confirmation email!

* *
Industry *
Work Phone Cell Phone


MIRRA - Cloud Resource Optimization


Recent Publications

MIRRA: Rule-Based Resource Management for Heterogeneous Real-Time Applications Running in Cloud Computing Infrastructures
Feedback Computing 2015;  April 14, 2015 

Automatic Resource Scaling for Medical Cyber-Physical Systems Running in Private Cloud Computing Architecture
Medical Cyber Physical Systems Workshop, Cyber-Physical Systems Week, 2014

An Auto-Scaling Mechanism for Virtual Resources to Support Mobile, Pervasive, Real-Time, Healthcare Applications in Cloud Computing
IEEE Network, 2013

Autonomic Computing Architecture for Real-Time Medical Application Running on Virtual Private Cloud Infrastructures
IEEE Real-Time Systems Symposium – WIP session, 2012



An assortment of question & answers to help you understand how this will maximize your profitability by saving you money!

How does MIRRA™ work?

MIRRA is middleware running in each virtual machine within the cloud environment to optimize resources such as the number of virtual machines, CPU sizes, etc. It supports public cloud solutions (e.g. Amazon Web Service, Google Cloud Platform, MS Azure, etc.) and private cloud solutions (e.g. OpenStack, CloudStack, etc.). MIRRA is mainly designed to support real-time machine to machine communication applications such as Internet of Things (IoT) devices, remote monitoring systems, etc.  MIRRA keeps a knowledge-base to adjust cloud resource size intelligently and automatically without human intervention.

Is MIRRA a stand-alone solution or does it require prerequisites?

It is a stand-alone and automatic solution. However, users should provide known requirements of their applications such as data transmission rate, data chunk size, application server addresses, etc. so we can understand your business logic for setting up the proper optimization rules.

How is Amerra Connect used with MIRRA? Is this something else that needs to be purchased?

MIRRA is a sub component of the Amerra Connect platform. Users can share the application key which can be used for both MIRRA and Amerra Connect. MIRRA can use Amerra Connect’s other components such as the key-value pair data store, file storage, user authentication, web dashboard, etc. without any restrictions.

Is MIRRA compatible with any device?

Yes, if the device needs to transmit data periodically like patient monitoring systems.

What are MIRRA’s potential limitations?

All or some of connected devices must be real-time applications transmitting time-sensitive data to their destined application servers in order to optimize cloud resources accurately.

Is MIRRA open source?

Currently MIRRA is not an open source project.

What language is MIRRA written in?

MIRRA is written in JAVA with Spring Framework and run as a system service in a virtual machine.

What is “System Reliability?

Processing real-time data is generally very time-sensitive for maintaining the minimum Quality of Service (QoS) because the current data must be processed before receiving the next incoming data. MIRRA provides a solution to observe and analyze the QoS requirements automatically and adjusts virtual resources including the number of application servers.

What is “Cloud Resource Optimization?

In order to serve all connected devices and applications, users must prepare any unexpected issues of service downgrades by running redundant application and database servers. Although, in an emergency situation, these redundant resources can be used to process real-time data, they are idle and waste their computing resources if other active services are being operated normally. This is because no quality of service requirement is provided by applications and devices. Since MIRRA can observe the quality of service requirements dynamically, it can minimize uptime of these idle resources. As a result of this adjustment, users’ cloud infrastructures can be optimized to serve their connected applications and devices.

What is “Real-time Scheduling?

Whether you use cloud infrastructures or not, the resource size of computing, networking, and storage services is often limited due to budget. If you have multiple running applications and devices, those must fairly share the resources needed to accomplish their jobs. If your applications and devices periodically send data to be processed by their application servers, a real-time scheduler can calculate an available job schedule for sharing the resources to maximize resource utilization without overflowing.

How much does MIRRA cost?

Although we provide a consulting service to integrate MIRRA into your existing system, it is essentially free to obtain. However, MIRRA requires to access the Amerra Connect platform to adjust resource sizes. Users must purchase the Amerra Connect license which can be charged per usage amounts such as the number of API calls. We are estimating $0.005/per API call from MIRRA.

How is MIRRA purchased?

For public infrastructure as a service (IaaS) cloud users who want to deploy their application servers, they can use our virtual machines registered in cloud market places such as AWS Marketplace, Google Apps Marketplace, Windows Azure Marketplace, etc. For private cloud users, they can download an installer from us, and configure their virtual machines manually. For Platform as a Service (PaaS) users, they can use our add-ons registered in Heroku, Engine Yard, Cloud Foundry, etc.

Does MIRRA require advanced software expertise to configure?

In order to configure MIRRA in your system, you should have enough knowledge to understand your application and device quality of service requirements. In the case that you have no software engineer or designer to do this, we provide a consulting service to help you setup MIRRA in your system.

Has this technology been tested?

We have conducted performance simulations with OpenStack which is a well-known private cloud infrastructure for remote ECG monitoring simulators. As a result of these simulations, we could increase around 80% of the system reliability and utilization.

How do you ensure confidential information, such as patient data isn’t compromised?

MIRRA must use the Amerra Connect platform and does not store any sensitive data internally. If it requires to store any sensitive data temporarily for MIRRA’s priority inversion process, Amerra Connect’s HIPAA compliant data store service can be used.

How are system reliability and optimization measured?

The system reliability can be measured by observing the real-time job drop rate. For example, your devices must send the certain number of jobs within the specified deadline, and your application servers also process them within another deadline. Even though your devices successfully transmit data, the application server might not be able to process them within the deadline without the real-time scheduler due to any resource shortage. Since MIRRA starts with running all the idle resources and optimizing their sizes by observing the systems, it can compare its performance to all other observing points.

The system optimization also can be measured by observing resource usages over time if the system reliability is kept as its minimum level.

Approximately how many IoT devices can be supported per virtual machine?

Since MIRRA does not provide any application server solution, it is impossible to approximate the number of devices that can be connected. It depends on how big data chunk is delivered, how long its transmission period is, and how quickly it can be processed by devices’ application server. MIRRA can calculate the number of devices per one application server after its learning period to optimize the cloud resource size.

Is this technology designed for all devices and if not, what is the ideal type of product that would benefit from MIRRA?

MIRRA is designed to support all device types if they are connected to virtual resources running in cloud infrastructures and transmitting time-sensitive data to their application servers periodically.

Is this technology needed immediately, or will it be needed in the future?

MIRRA is designed to support all device types if they are connected to virtual resources running in cloud infrastructures and transmitting time-sensitive data to their application servers periodically.

Why are you targeting healthcare first? Can this be applied to any industry?

Most types of medical IoT, wearable, and patient monitoring devices require real-time transmission of sampled data to their application servers. These jobs are extremely time-sensitive for transmitting a patient’s health condition. MIRRA is perfect for this extreme case but can be applied to other cases monitoring non-human objects such as gas pipelines.

How does this compare to VM ware & vCloud?

VM ware is a hardware virtualization solution which can be used for vCloud, a private cloud solution using VMware. MIRRA can be used by users of vCloud.

How does MIRRA differ from existing auto-scaling features from cloud infrastructure companies?

Currently used auto-scaling features are based on rules related to a virtual machine’s performance metrics which are not directly related to application server’s performance. Since these rules are determined by system administrators before initiating the entire system and updated manually too, these auto-scaling features are used only for preparing cloud resources preparing the worst case scenarios keeping all idle resources.

Briefly, who are the immediate competitors and what separates MIRRA from them?

Cloud orchestration companies supporting auto-scaling features can be considered competitors such as RightScale. However, all still provide the performance metric-based auto-scaling features.

Is MIRRA patented? If not, are there plans to or is it needed?

Its algorithm can be patented, however, penetrating the market and having more users are much more important during this initial go-to-market strategy.