Tools

GHOST – Gateway Host Nodes – Distributed Integration Platform

When organizations want to engage in digital transformation, a traditional hub-spoke approach (central) fail to deliver necessary implementation speed, is significantly driving cost and will become difficult to maintain.

A distributed architecture is a prerequisite to support innovation and an explorative rapid development style. This can also ensure low cost when scaling and support reliable end-to-end operational monitoring and control capabilities.

The GHOST integration platform is fully distributed and has proven capabilities as the key component in high frequency data collection and aggregation in an Internet of Things (IoT) approach. The GHOST platform can also handle classical integration scenarios between applications within an organization while keeping operational cost and implementation effort on very low level.

As new requirement arises on classical integration, GHOST is also capable to handle integration scenarios between partners, with cloud services and between cloud services.

Why was GHOST invented?

Ghost Labs, in collaboration with the Think Tank Duqtor, Stockholm University and the Royal Institute of Technology, identified that current integration methods have a major obstacle to manage new integrations meeting the demands of Digitalization and Industrial Internet of Things (IIoT) scenarios.

Based on this, a distributed integration platform optimized for IIoT was developed, that has been proved to significantly reduce cost, lead-time and complexity while simplifying scalability, improving reliability and robustness.

The close cooperation with research institutions has ensured the integration platform is built per the latest research available. In late 2014 the GHOST integration platform made it to the final and top 10 out of 163 competitors in a EU wide contest for innovators in the category IoT.

GHOST can be used as the framework to create integrations and setup of nodes collecting data in a distributed landscape (also including partner and cloud environments). These nodes are key in the GHOST setup. All integrations and logic are configured in a centrally development and control tool.

From this central tool, the nodes are then automatically pushed out in the distributed landscape as close as possible to the actual data sources.

This allows point-to-point data flows that ensure the shortest and most efficient data transfer between data sources in any distributed architecture.

In addition to transport capabilities, the nodes include agents (a small piece of software) that hold algorithms for managing translation, aggregation and any business rule and logic.

This minimizes incidents since the data is only translated once and is flowing point-to-point with the relevant data only.

The nodes and agents ensure that data and information are processed and distributed to the right recipient in a network similar to the principles used in the Internet protocol.

Agents in the nodes can hold business rules and can easily be configured through an open API. This allows GHOST to be included and integrated in IoT applications. GHOST will be the engine in classical and new integration scenario setups, replacing hub-spoke (central) architectures.

GHOST is designed and priced to fit small, mid-size and large companies.

The GHOST platform

Distributed vs. A Central Integration Solution

Integration is an exchange of information between different actors with an interpreter who translates from one format to another. In a hub-spoke approach, typically seen in an Enterprise Service Bus (ESB) setup, the ESB become a single point of failure for all data transfers in addition to a point-to-point solution.

To ensure reliability and according to security, performance and up time requirements the cost for running an ESB becomes very expensive compared to GHOST due to high licenses and hardware costs.

With a distributed solution, the data transfer can follow the shortest route and leverage existing hardware that the data sources already are running on.

It also significantly increases traceability compared to a traditional hub-spoke approach, as it allows for true end-to-end visibility within the control span of the data transfer, all visible in the GHOST monitoring dashboard.

Central Governance

Even if GHOST integrations is set-up as point-to-point transfers, the central management and control tool ensures a single point of management and monitoring.

This is why hub-spoke approaches became so popular when introduced 15 years ago, as it allowed for better central governance and control of the integration landscape.

GHOST offers a very efficient and low cost point-to-point integration, as well as full central control and configuration. All in an intuitive tool with a user interface that simplifies configuration and support.

The use of GHOST has showed that design and configuration is more than 30% faster than in a traditional hub-spoke approach.

Another very strong feature with the agents besides that they execute translation, business rules and logic next to the nodes, is that they are also configured from and pushed out from the central GHOST configuration tool.

With GHOST it is very simple to reconfigure e.g. data collection frequency and processing of data at the very source, managed from a central tool. This is especially important in IoT solution and setup.

State of the art Architecture

The distributed and network-based architecture is ”state of the art” in which the distributed solution will continue to function even if parts of the network or the system malfunction. The solution runs entirely in the cloud, within the company’s or partner networks, or as a hybrid.

The logic of GHOST

There are already a large number of plugins (APIs) out of the box in GHOST. An unlimited number of Nodes can be set up in the solution.

GHOST monitors each steps of the integration chain with alarms when there is a deviance from the predetermined thresholds. This means that the failure of the integration chain can quickly be found and notify the right people with the right skills for the job or secure e.g. an automatic shutdown of a machine in real time.

IIoT – Industrial Internet of Things

Industrial Internet of Things (IIoT) solutions require completely different design principles and architecture.

GHOST is built to handle that.

In an IIoT solution an extreme amount of information can be generated, hence the system cannot be allowed to malfunction. A sensor or probe can produce hundreds of signals per second and a machine or system can have many sensors or probes that generates a huge amount of information and requires that the agents is as close to the data source as possible.

The agents, with its logic, ensure that the information distributed is relevant. Traditional solutions do not support this design pattern and logic.

Attractive Total Cost of Ownership (TCO)

The pricing model for GHOST is based on the actual use, i.e. no traditional license fees.

Due to the small footprint you can easily use the existing hardware for the GHOST platform. The inbuilt reliable and secure data transfer notably reduces the total cost of ownership (TCO) compared to a more complex hub-spoke solution.

With the GHOST approach, TCO can be reduced by 50% in addition to more than 30% efficiency gain in design and configuration.

This together drives down the cost significantly.

Benefits going from a centralized to a distributed integration setup in a large industry company show:

• Costs reduced > 70%

• At the same time Gartner considered 35% as a benchmark

• Agility/flexibility increased > 30%

• This was possible both as integrations are more direct leading to reduced complexity, and as reuse and configuration take less time than coding
• Reduced the share of integration work in new integration projects from 35% in 2002 to 15% in 2012

(Case study with the use of GHOST at a large manufacturing industry)

Principle of Design and Who implements GHOST?

GHOST can easily be managed by companies and organizations themselves but also accompanied by implementation partners.

Working with a solution like GHOST can have a major impact on the business. So when you design a solution with a new approach to architecture this might even change the current business model, you have to:

1. Focus on the big picture
2. Create commitment for all the individuals involved, from management to developers
3. Create an exploration/experimental approach.

There are also methods available that can support you in this.

Support

Product Support is managed by Ghost Labs for GHOST and Ghost Lab developed agents and plugins.

The Implementation Partner has a responsibility to support the implemented solution.

 

For more information and to apply for a GHOST test access, please contact:              

Lars Bengtsson at +46 70 279 33 29   or lars.bengtsson.swe@gmail.com

 

Example of relevant use cases suitable for GHOST

1. Any industry
   1. Replace expensive FTP solutions
   2. Replace expensive legacy integration solutions
   3. Collect log files from critical systems and notify actions both locally and centrally
   4. IoT or IIoT initiatives
2. Manufacturing
   1. IIOT – Use GHOST as a backbone for efficient use of sensor data in production.
3. Finance and insurance
   1. Blockchains – GHOST can distribute critical information to the right source and distribute correct agreements to a blockchain actor.
4. Utilities
   1. Management of Smart Home data.
5. Telco
   1. Eliminate Information islands. E.g. distribution of relevant transmission information stored at a network cell
6. Supply chain
   1. g. exchange information with distribution hubs or carriers
7. Automotive Industry
   1. Collection of vehicle data
8. Public sector
   1. Secure information flows in multi agency’s processes
   2. Use ghost for distribution of documents

In order to create the final solution you might need to combine other tools of choice for visualizing the information or process the information e.g. BI, AI etc.

This are examples only, GHOST can replace traditional Integration Solutions in many other use cases.

Summary GHOST

GHOST will be an important part of your environment; GHOST is the component responsible for the exchange of information like logs files or documents, critical information to and from systems or from sensors and probes.

1. Substantial reduction of integration costs
   1. A very attractive pricing model where you only pay for actual use.
   2. Reduction of complexity makes you much more agile and flexible.
   3. The distributed architecture make the cost of incidents becomes lower due to the fact that there are fewer sources of error:
      1. It is a ”simple” solution
      2. Fewer actors involved

• Incident management becomes much faster, one can immediately see where the error occurred.

2. The solution is prepared to handle ”Things” (IIoT or IoT).
   1. Through the use of agents, close to the nodes, consisting of algorithms deciding what information is relevant to redistribute, it minimizes bandwidth which can become a problem when, for example sensors can emit hundreds of measurements per second.
   2. Some incidents require immediate action, for example on a machine level. E.g. a drill may need to be stopped if it vibrates too much. While the relevant measurement values are distributed to the next node for the compilation of valuably business information.
3. Easy to use
   1. Intuitive user interface.
   2. Automatic ”deployment”, updates and instrumentation
   3. Prepared for mobile use.
   4. Extremely scalable
4. Hybrid
   1. On-demand, Roaming and up-to-date
   2. In the cloud, on premise or hybrid
   3. Suitable for all types of businesses, small, medium size and large.
5. Distributed Network
   1. Information travels the shortest route with a minimum of actors involved
   2. Infinitely scalable in the cloud or as hybrid.
   3. The strength of a distributed system landscape.
6. Central governance
   1. Central monitoring and deviation management
   2. Multiple views of your integrations
      1. Technical
      2. Domain
   3. Notifications can be sent to any tool of choice.