Virtual Runway Project

Virtual Runway Project
You just need focus on part 4 and 5 of this project that in red words, because this is a group project. You should have a whole look of the whole paper before you
start work on part 4 and 5! You need increase these two part to 2 page in length. For the resource, you can refer the reference at the end of the draft and the
resource at the end of the requirement, or you can search online. And the first attachment is the requirement of the project and the second one is the draft which we
already finished.

Virtual Runway Project
Technology Requirements

Objective
This deliverable presents the requirements technology analysis for the Virtual Runway project. The outcome of the analysis is a consolidated list of requirements that
define the project’s goals and that will guide the Virtual Runway technical activities. The rational behind these requirements is to develop appropriate methods and
tools matching industry real-life needs.

A critical success factor in the development of high quality software applications is a deep understanding of the projects’ requirements, both technological and
business user needs. This requirements document will serve as a baseline of the analysis process, so as to ensure that all of Virtual Runway’s interests will be
completely covered and also to avoid requirements that are out of scope. Hence, the resulting requirements specification envisions the foundations for a successful
implementation of a Virtual Runway prototype and for creating applications that satisfy the real business needs.

Business Case

As the technology enabling virtual reality (VR) and augmented reality (AR) continues to improve and become more cost effective, opportunities are arising for retailers
to revolutionize the way that consumers preview fashion and buy clothes. No longer will it be necessary to drive to a store, navigate crowds, find suitable products
(with or without the assistance of knowledgeable salespeople), and try on the clothes before deciding which ones to purchase. Instead, this entire process could be
digitized and have the added benefit of improved customer service and the ability to create customized products based on individual preferences. Such a marked
improvement in the customer buying experience will likely lead to enhanced loyalty, repeat sales and greater revenues. Although still in its infancy, this technology
will empower the “stores of the future,” and retailers, especially clothing boutiques providing the latest high-end fashions, are well advised to develop their
capabilities in this space before their competitors are able to gain the advantage.

Risks

Physical Risk
• motion sickness, dizziness, epileptic seizures, blackouts, triggers phobias

Security Risk
• ability to modify appearance of avatars to impersonate other users
• ability for users to modify the environment itself
• Storing multiple high resolution customer images, including potentially those of minors and children, presents unique challenges.
• The product will need encryption, stringent protocols and constant intrusion detection.
• A cloud-based solution, with the cloud provider handling security, could be a cost-effective solution that requires little ramp time

Behavioral Risk
• potential for rudeness, harassment, stalking (same risks as in a physical environment)

Privacy Risk
• every behavior in virtual environment can be tracked and every element manipulated
• violations create potential problems with FTC and state regulators
• violations of employee privacy rights have legal consequences

Cost Risk
• The hardware cost will scale rapidly if the product is successful.
• Even if the storage and bandwidth per customer scale linearly (X MB/Customer and Y MB/Second per Customer), a rapid increase in the number of customers could
overwhelm us.
• Separately, there are substantial capital costs to the system – showroom hardware, data center storage, software licensing costs, custom software development
etc. To mitigate this risk, we suggest starting with a prototype system in as few stores as possible. Based on the appeal of the product and the lessons learned from
the prototyping, the final product could be refined and deployed in a production environment with a cloud provider.
• A cloud service could, at a minimum, provide cost certainty (and revenue requirement visibility) on the cost per customer scaling.

• Prototyping has an additional benefit – it provides some visibility into the ROI of the system. If stores with the system show, ceteris paribus, meaningfully
higher sales than stores without the system, we can attribute the incremental revenue to the system. That would allow us to calculate an order of magnitude ROI.

Technology
Given that much of the current technology is more suited for gaming than for business, the Virtual Runway design will regularly undergo technology consistency checks
to make sure the design supports all new or evolving VR technologies. Said technology consistency checks will take place every time Virtual Runway technological
enhancements are proposed and discussed.

• slow adoption of technology by target market

2. Technical Design Diagram
Virtual Runaway Design Diagram:
FIG. 1 a block diagram illustrating the virtual reality shopping experience in accordance with an embodiment of the present invention.

Suggestion: made as a diagram
3. Database examinations (which ones and why)

By utilizing VR technology, the data collected from customers and produced in analytical process will definitely boom. According to our business requirement, the data
sources that will be involved in analysis are structured data, unstructured text data, unstructured image data, and real-time data. The retail store will require about
0.5 petabytes of data per month and 10 million filing cabinets worth of text data per month. So how to effectively manage such large volume and various type of data?
The key is to introduce professional database framework engine into our company. Currently, Hadoop, Spark, and NoSQL are the most widely used database system in
practice, so we will take these three into account.
Considering our company’s size and its sales volume, we will use hadoop as our start database engine.

Hadoop:
Hadoop, as a major platform to differentiate useful dataset from tons of data, can efficiently retrieve insightful information. Meanwhile, it is designed to scale up
from single server to thousands of machines which offer individually computation and storage (Carey, 2016). However, we still need to comprehensively evaluate its pros
and cons, then make decisions.

As for pros, first and foremost, Hadoop can deal with large range of data sources. Since the data of virtual reality retail business could be structured, like age and
income, unstructured, like text as well as image, and even real-time, so Hadoop is an ideal choice to handle such complex situation and save processing time. Then,
Hadoop is also cost-effective (KnowledgeHut, 2016), which means that it can store the entire raw data produced in whole business and prevent losing valuable
information. It provides an easy way for virtual runway to get access to the historical data and without concerns about extra expenses. Last but not the least, Hadoop
could keep a high speed of storing information when consumers are browsing the items and virtually try out, which is essential to identify their information and
provide proper recommendations.

In terms of cons, the biggest trouble is security issue. On one hand, Hadoop is one framework built entirely on Java, which is most widely used programming language.
Thus, the platform is kind of vulnerable and exposed to the danger of cyber criminals. On the other hand, in Hadoop, the security measures are inadequate, so it
requires specific team responsible for developing securing actions.

Overall, pros obviously outweigh the cons. Due to the relatively low costs and the trait of open source platform (Carey, 2016), Hadoop will be the prime choice for our
company with a booming data pool. When the database is mature and funds are adequate, we could consider moving to NoSQL and Spark.

NoSQL:
NoSQL, which is the collection of databases and does not use SQL or other relational data model, is used for rapid data storage, increasing the simplicity of data
concurrency and integration. The four basic databases for NoSQL are Key-Value Store, Column-Oriented Store, Document Store, and Graph Databases. They are all very fast
and can be easily dealt with different types of database. Since NoSQL is low cost and is widely used for non-relational database, it will be a good choice for our
company to use this tool for data storage. If the model is easy, needs flexibility and does not need a high degree of consistency, the NoSQL would be a great choice
for us applying to a VRun system. NoSQL can deal with huge data base, and its horizontal scalability would enable people easily have the relational database system. We
can also use NoSQL to store the data source such as the unstructured image data and UX (users’ experience).
Besides its convenience and simplicity, there is something we need to consider, such as data security, governance and limitations. HBase which is one of the database
of NoSQL has a complete governance and security system while others does not have. This would be vital for a retail store that the security of their data storage. It
is related to their responsibility and credibility to customers. This will be an important thing when we are using NoSQL as storage tool.

Overall, the NoSQL database will be a good choice of data storage when we are starting to use VRun technology into retail store since they are flexible, elastic
scalability and high performance, but we may also focus more to the data security and combine it with more developed technology such as Spark.

Spark:
As our sales developed and widely using VRun in our stores, we can transform from hadoop to Spark. Spark, which is a cluster computing framework, has been widely used
in streaming data analytics, graph analytics, machine learning and etc. Its main design concept is to develop a system that is faster and easier for analytics. For
instance, some of the main features of the Spark are that it runs 100 times faster than MapReduce, many functions and operators available for data analysis, can
process iterative and interactive analytics and etc. For our retail store, there are numerous kinds and sizes of data. Spark would be one of the engine that makes the
processing easier and faster.
Also as mentioned in our first database engine, Hadoop, Spark would be one of the most partner cooperating with it. Hadoop, which only contains map and reduce, need
higher level of coding skills and phases. However, in Spark, this has been solved. It is also provided with many libraries, Spark SQL, Graph X, MLib, and Spark
streaming. Spark deals with database much quicker than Hadoop, and this will save a lot of time in data processing. These libraries make spark highly optimized for
general use. Besides its convenience and simplicity, it is memory hungry. In other words, it needs memory, and would take large resources.
For the final stage, we probably will introduce Spark to our database engine.
4. Functions / Functional Detail
Data functions
• Insert Large files
• Storage Information

File functions
• Hadoop’s HDFS would be one thought

Web functions (if need be)
Application functions
• Writing fast
• Doing multiple jobs

5. Access Requirements

• Analytics Team and Technical people are required to have this access
• Developer will have the full access
• Marketing Team will have the data access

6. Storage Needs
The storage needs will scale rapidly if the product Is a success and is adopted by multiple vendors. It is easy to see petabyte-scale storage. Therefore, the
deployed system must be able to scale in a cost-effective manner. This favors Hadoop. Given the potentially significant infrastructure costs, we are investigating a
cloud-based solution That shifts hardware responsibility to the cloud vendor.

7. Resource Needs
Given the 15 second rendering business requirement, the product has significant bandwidth requirements. More specifically, it needs to handle both large volumes of
data (images) AND have low latency. As such, we are considering a geographically distributed architecture that would place the rendering engines close to the
showrooms (and, for the virtual showroom, at multiple internet hubs across the country) and the data at multiple data centers across the country. This would allow the
showrooms (and the virtual showroom) to quickly render images using local data or data from (relatively speaking) nearby data centers. The prohibitive costs of
leasing and stocking so much hardware make a cloud-based solution more attractive. Hadoop with mirrored data may be a potential solution.
8. Costs (broadly)
• price tags for custom virtual environments can easily climb into 6 digits and higher
• hard to determine ROI
9.References

Knowledgehut. (2016, November 25). Top Pros and Cons of Hadoop. KnowledgeHut Blog. The link: www.knowledgehut.com/blog/bigdata-hadoop/top-pros-and-cons-of-hadoop

Scott C. (2016, July 5). Hadoopvs Spark: Which is right for your business? Pros and cons, vendors, customers, and use cases. ComputerworldUK website. The link:
www.computerworlduk.com/data/hadoop-vs-spark-which-is-right-for-your-business-pros-cons-vendors-customers-use-cases-3642888/

Order from us and get better grades. We are the service you have been looking for.