This document discusses a project on developing a context-aware computing infrastructure for VMware ecosystems. It provides background on VMware, prerequisites like context-aware computing and virtualization. It then outlines the project to develop a prototype application that gives dynamic insights for effective power usage and virtual machine/database placement in data centers to reduce costs. Challenges included mapping context awareness to cloud/data centers and finding a solution for power management and application placement. The timeline shows progress from researching context-aware computing to designing and mapping a solution.
Context-Aware Computing Optimizes VMware Data Centers
1. Context Aware Computing infrastructure
for VMware ecosystem
Shiva DS
Technology Exploration and Prototyping
2. Thanks to
Head of the Department:
Dr. R. Nadarajan,
Professor & Head,
Department of Applied Mathematics and Computational Sciences,
PSG College of Technology, Coimbatore.
Program Coordinator:
Dr. N Geetha,
Program Coordinator,
Department of Applied Mathematics and Computational Sciences,
PSG College of Technology, Coimbatore.
3. Thanks to
Tutor:
Dr.N Mohanraj,
Associate Professor,
Department of Applied Mathematics and Computational Sciences,
PSG College of Technology, Coimbatore.
Internal Guide:
Dr.N K Sreeja,
Assistant Professor,
Department of Applied Mathematics and Computational Sciences,
PSG College of Technology, Coimbatore.
4. Thanks to
Manager and Academic Guide:
Mr. Sairam Veeraswamy,
Senior Director,
Innovation Programs - India R&D,
VMware,
Bengaluru.
6. About VMware
•Vmware was formed in 1998, Palo Alto, California, United
States.
•VMware, Inc. is a subsidiary of Dell Technologies that provides
cloud computing and platform virtualization software and
services.
• It was the first commercially successful company to virtualize
the x86 architecture.
•The first product, VMware Workstation, was delivered in May
1999.
9. Context Aware Computing
•It is defined as the “software that examines and reacts to an
individual’s changing context”.
•Context characterizes the situation of an entity.
•Google maps, Pokemon Go, Air Conditioners,
Refrigerators, Facebook and Cardiac Monitoring.
23. My Project
•Mapping out the competitive landscape.
•Identifying cloud computing and data centric use cases.
•Developing an intelligent application prototype which gives
dynamic insights for,
Effective usage of power within data center.
Placement of VMs and DBs during application deployment
to increase the ROI for customers by decreasing traffic.
29. Migration
•Migration allows a VM to be moved from one server to
another so that the server can be switched off or moved to
power saving mode.
•Identifying the VM on the server with low utilization that
could be migrated, so that the provider can put servers on
the idle or power off.
30. Power-On/Off Servers
• Switch-off unused servers.
• Predict the incoming requests.
• Switch-on required servers on the arrival of new requests.
31. Prediction based algorithm
•Involves scheduling and prediction mechanism to efficiently
use the green energy sources.
•Solar and smart grid is the major power source for data
center.
•Predicting the availability of solar energy.
32. Cost Aware
•Taking advantage of different electricity prices in different
geographical locations.
•Its objective is to reduce energy cost while assuring delay
constraints.
•This approach is not applicable for interactive jobs and web
requests that are time sensitive.
33. Location
•Maintaining the amount of green energy produced across
all the data centers.
•Migrating the VMs across the data centers whenever the
green energy produced locally becomes insufficient.
•Network delay and the amount of energy consumed for
migration must be taken into considerations.
34. Heat Management
•Nearly 45% of the power are used in cooling in data
centers.
•Server generates a lot of heat when the workload is huge.
•Maintaining a critical temperature for cooling.
36. Problem
•A huge amount of traffic is generated inside data center.
•Applications usually have multiple computing(VM) and
associated data components.
•Some of the multi-component applications are social networking
sites, ecommerce and Internet-based gaming applications.
•The computing components are delivered in the form of VMs
whereas data components are delivered as data blocks.
37. Network Data Aware Placement
•Mapping of computing and data components of the application
into the computing and storage nodes of the data center.
•The computing, storage and network constraints must be
satisfied.
•NDAP strives to reduce the distance that data packets.
•This reduces the localized network traffic and reduces overhead
in the upper layer network switches.
38.
39. Challenges faced
•Mapping context awareness to cloud and data centers.
•Generating use cases.
•Mapping out the competitive landscape.
•Finding a solution for power management and application placement in
data centers.
40. July 1st
week
• Report on
CAC
July
2nd
week
• Use cases
July 3rd
week
• Elaboration
of use cases
July 4th
week
• Competitors
architecture
Aug 1st
week
• Project
design
Aug 2nd
week
• Mapping
solution
Timeline