Due Acknowledgement of the Reference URL at:
http://jineshvaria.s3.amazonaws.com/public/cloudarchitectures -varia.pdf
https://developer.apple.com/machine-learning/
https
/
://cloud.google.com/ml
engine
An Overview
Introduction to Cloud Architectures & Computing objectives of cloud architectures/platform, applications, etc;
Lesson Learnt from the GrepTheWeb Tips for developing effective cloud applications;
More detailed considerations of the concepts/factors involved;
Summary
Introduction to Cloud Architectures & Computing
Basically, different vendors/research groups may provided their own definition of cloud architectures and computing;
In the subsequent discussion, we focus mainly on the GreptheWeb system, as built by the Amazon group, as an example of the Cloud architectures & applications for the lessons they learned.
The Motivation Behind the Cloud Architectures & Computing
Cloud architectures address several key issues on large-scaledataprocessing, real-timeand/orresource- intensive computations as required in complex analyses or models of industrial (or even mission- critical) applications;
Basically, applications built on Cloud architectures will run in-the-cloud where the physical location of the infrastructure is determined by the service provider, mostly using the on-demand basis;
Business Benefits of Cloud Architectures
Almost zero upfront infrastructure investment: with the utility-style computing, there is NO fixed or startup cost;
Just-in-time infrastructure: cloud architectures can relinquish the required infrastructure and services as quickly as you requested and got them in the first place;
More efficient resource utilization: with cloud architectures, one can manage resources more effectively and efficiently by having the application & resource requests on-demand;
Examples of Cloud Architectures
There are numerous examples of cloud architectures including:
Processing pipelines
Document processing pipelines to convert hundreds of thousands of documents from one format to another, e.g. from MS Word to PDF;
Image processing piplelines say resizing millions of images;
Video transcoding piplelines transcoding AVI to MPEG movies;
Indexing creating an index of web crawled documents;
Data mining performing search over millions of records.
Batch processing systems: back-office applications in finance, insurance or retail sectors, log analysis, nightly builds or automated unit testing and deployment, etc.
Examples of Cloud Architectures(2)
Websites
Websites that sleep at night and auto-scale during the day, e.g. to analyse or back up the transactions performed during day-time;
Instant websites websites for conferences or sports event (such as the Super Bowl, sports tournaments);
Promotion/Seasonable websites that run only during the tax season or the holiday season (Black Friday or Christmas as in US).
An Example of Cloud Architectures: GrepTheWeb
The application is in production at Amazon.com with the code name as GrepTheWeb simply because it can grep (as a popular Unix command to search patterns) the actual web documents (say in millions of crawled documents);
GrepTheWeb allows developers to perform fine- grained & specialized searches like selecting web documents containing a particular HTML or META tag, finding documents with particular punctuations, searching for mathematical equations like f(x) = x + W, source code, email addresses or other patterns.
Level-1 (Simple) Architecture of the GrepTheWeb Application.
The GrepTheWeb application uses highly-scalable components of the Amazon Web Services infrastructure that not only scale on-demand, but also are charged for on-demand.
Level-2 (more detailed) Arch. of the GrepTheWeb Application..
The 4 Phases of the GrepTheWeb Application..
Detailed Explanation of 4 Phases
Launch phase: validating and initiating the processing of a GrepTheWeb request, instantiating the Amazon EC2 instance, launching the Hadoop cluster and starting all job processes;
Monitor phase: monitoring the EC2 cluster, maps, and checking for success or failure of the jobs;
Shutdown phase: for billing and shutdown of all involved processes and the EC2 cluster;
Cleanup phase: deletes the transient data stored in the Amazon Simple-DB (database) storing all the inputted documents.
Essential Concepts Behind the Design of GrepTheWeb App
Similar to the loosely coupled and distributed reservation stations as in Tomasulo !
Lessons Learnt from the GrepTheWeb Application
FIVE TIPSfordevelopingeffectivecloudarchitecture applications:
1. Ensure that your application is scalable by designing each component to be scalable on its own. If every component implements a service interface, responsible for its own scalability in all appropriate dimensions, the overall system will have a scalable base.
2. For better manageability and high-availability, make sure that your components are loosely coupled. The key is to build components without having tight dependencies between each other, so that if one component were to die (fail), sleep (not respond) or remain busy (slow to respond) for some reason, the other components in the system are built so as to continue to work as if no failure is happening.
Lessons Learnt from the
GrepTheWeb Application (Contd)
3. Implement parallelization for better use of the infrastructure and for performance. Distributing the tasks on multiple machines, multithreading your requests and effective aggregation of results obtained in parallel are some of the techniques that help exploit the infrastructure.
4. After designing the basic functionality, ask the question what if this fails? Use techniques and approaches that will ensure resilience. If any component fails (and failures happen all the time), the system should automatically alert, failover, and re-sync back to the last known state as if nothing had failed.
5. Do not forget the cost factor. The key to building a cost- effective application is using on-demand resources in your design. Its wasteful to pay for infrastructure that is sitting idle.
Cloud Servers Supported by GPU
Computing
Most modern cloud servers, such as the Google Cloud, Apples iCloud or the Amazon EC2, are often well supported by the backend GPU clusters to provide highly efficient and scalable computing power for A.I., big data analytics or machine learning tasks;
For instance, the
Engine (URL : https://cloud.google.com/ml
Google Cloud Machine Learning (ML)
isa managed service enabling developers and data scientists to
engine
/
)
build and bring superior machine learning models. Cloud ML Engine offers training and prediction services, that have been used by enterprises to solve problems ranging
identifying clouds in satellite images, ensuring food safety, and responding four times faster to customer emails.
from
Cloud Servers Supported by GPU
Computing(2)
The Apple iCloud provides the Core ML 2 (URL : https://developer.apple.com/machine-learning/) that allows fastperformancewitheasyintegrationof machine learning models thru the Create ML and playground in Xcode 10;
The Core ML 2 supports many standard ML models such as tree ensembles, SVMs, etc.
On top of the Vision Framework includes interesting features like face tracking, object tracking, etc. while the Natural Language Framework is a new framework to analyze natural language text and deduce language- specific metadata.
Reviews
There are no reviews yet.