The Stream Data Processor is designed to perform complex queries and calculate different KPIs based on dense data streams of online customer behavior performance counters.
Apache is an XML and SOAP object model which supports deferred parsing and on-demand building of the object tree.
You can extend the capabilities of your web services with our .NET Framework technologies.
We offer support for building XML Web services, using technologies designed to meet the needs of different audiences.
Maana search and knowledge engine for big data on Hadoop addresses an immediate and growing fundamental need at the world’s largest 3,000 corporations and government agencies.
DEMAND: Enterprises are opening up their big data to employees & customers. They require a horizontal search platform on their big data to be able to do so.
OPPORTUNITY: Incumbent enterprise search engines, in relation to the nature of the task at hand on big data, pose prohibitive storage & compute costs and come with design limitations.
PROBLEM: The only alternative that is available to enterprises today is to attempt to devise their own in-house search platform on their big data through Hadoop.
PAIN: The labor required is not trivial. Enterprises lack the deep expertise needed. Also, the foundation they thought they can adopt and build on top of, Lucene, has severe shortcomings.
FEEDBACK: We spoke with 51 of the 377 Fortune 1000 corps that have Hadoop. 40 are opening up their data and face this problem, while 33 are experiencing the pain.
MAANA: Enterprises will be able to do on their big data what Google Knowledge Web and Bing Satori did in their walled-gardens. Maana will be plug & playable for most of our customers.
CORE IP: In-memory emergent-semantic-graph representation of knowledge and the extensible-semantic-processing-pipeline that assures lowest compute costs beyond petascale.
TARGET MARKET: The world’s largest 3,000 corporations and government entities. 20% have Hadoop deployed in production. We expect 90% by the end 2016.
IMMEDIATE MARKET: 90 Fortune 1000 corps that centralized the management of Hadoop. This transition occurs with 2-3 years of first (departmental) Hadoop deployment.
MARKET TRACTION: 6 Fortune 500 corporations are considering to be our early adopters, 3 of which, GE, Samsung and Intel, are also evaluating investment at our current Seed B Round.
With the release of a new version, AWS Identity and Access Management (IAM) began to support identity federation for delegated access to the AWS Management Console or AWS APIs. Now the external identities (federated users) are granted secure access to resources in enterprise’s AWS account without having to create IAM users for each individual user.
With the need to establish a relation between corporate users and IAM accounts on a corporate level, Akvelon was tasked to provide more secure possession and use of IAM accounts by the enterprises to which these IAM accounts were issued (their identities and secret keys are never shared with end users). We were also asked to provide the ability to quickly adjust federated corporate user roles/permissions by simple action at the Active Directory level.
Akvelon was chosen as the sole contractor for this project simply due to the aggressive timeline Akvelon agreed on, and the fact that Akvelon proactively proposed several viable solutions at the stage of initial discussion. Akvelon assumed full responsibility end-to-end for this fixed-bid contract.
The Identity Federation application is an ASP.NET MVC web application hosted on corporation premises.
From a functionality perspective, it is a proxy, which determines the Windows identity and group of the user through Active Directory. The application creates a proper request to Amazon Identity and Access Management services, containing an AWS Security Token Service (STS) for proper authentication and authorization on Amazon’s side.
The application was supplied with Windows Installer, allowing for rapid enterprise deployment. Amazon decided to ship the application with open source code, allowing the IT department of each enterprise to adjust federation process if they chose to; or just use the application as developed. The final deliverable was published by Amazon.
The main concern was to design software that would be able to process large data volumes in real-time mode with a reasonable value of Total Cost of Ownership. This meant that classical, pure database-based solutions would be too expensive in acquisition and maintenance. Nevertheless, Akvelon managed to design a complex system with an attractive cost-effectiveness ratio.
We designed and developed the solution based on five major software components:
Stream Data Processor: The Stream Data Processor is designed to perform complex queries and calculate different KPIs based on dense data streams of online customer behavior performance counters. For the Newest Complex Event Processing (CEP) engine, Microsoft StreamInsight was used as a technological platform. It allows analyzing and correlating data incrementally while the data is in-flight, which yields very low latency and eliminates the need to have a complex and expensive data warehouse. Only meaningful data (aggregates, calculated KPIs values, etc) are subject to saving in the persistent data storage. Another benefit is that the solution can be quickly and effortlessly adjusted to respond to areas of opportunity or threat by incorporating new KPI definitions into the logic of the application, thereby improving operational efficiency and ability to respond quickly to business opportunities.
Data Storage Subsystem: The Data Storage Subsystem is designed to store historical data, and it uses the Stream Data Processor module as a data source. The module implementation relies on power and flexibility from the Microsoft SQL Server relational database.
Analytics Product: The Analytics Product is designed to monitor statistics and run various types of KPI related reports. The module’s user interface was implemented as a lightweight browser-based solution and requires no additional software installed. This component has a custom interface to meet beginner and experienced user requirements.
Notification Product: The Notification Product provides a scalable and flexible mechanism for delivering notification messages (e-mail, SMS) based on predefined KPI related predicates. The notification engine evaluates predicates every time a KPI value is changed. The engine is triggered according to easy, customizable, and flexible rules.
Admin Product: The Admin Product is developed to supervise and, if necessary, take control of all activities in the system.