1、Introducing Cloud Computing into STEM Curriculum Using Microsoft Azure,B. Ramamurthy Computer Science and Engineering Dept. University at Buffalo binabuffalo.edu http:/www.cse.buffalo.edu/faculty/bina,6/2/2011,Cloud Futures 2011, Redmond, WA,1,Introduction: A Golden Era in Computing,6/2/2011,Cloud F
2、utures 2011, Redmond, WA,2,Topics for Discussion,People involved in this project Challenges in introducing Cloud Computing Educational imperative Efficacy of existing methods Cloud models General approach introducing cloud computing to STEM curriculum A Transformative Pedagogy Demo 1: Name2Face : Ad
3、dressing CSE audience Demo 2: Digital Biology: Addressing broader STEM audience Features of Azure and Challenges Summary,6/2/2011,Cloud Futures 2011, Redmond, WA,3,People on this project,This is a work in progress. Damian: an undergraduate sophomore honors student who chose to work on this project a
4、s his summer research Suchismit: a graduate student with deep theoretical background who wants work with cloud computing research Ying: an enthusiastic female graduate student who likes to implement projects on the cloud,6/2/2011,Cloud Futures 2011, Redmond, WA,4,Challenges in introducing Cloud Comp
5、uting,More people are familiar with cloud computing when compared to the Internet during its infancy Educators as well as students want to learn and use the “cloud” and “cloud computing” However, CSE and STEM (Science, Technology, Engineering and Mathematics) curriculum is quite rigid steeped in tra
6、dition and standards Current curriculum is so overloaded there is very little opportunity for new material Server-side and large-scale application development is indeed a challenge with storage needs, ports, web servers, firewall and vulnerabilities Educators are overwhelmed preparing for the dense
7、content of the existing curriculum, newer concepts take a backstage,6/2/2011,Cloud Futures 2011, Redmond, WA,5,Educational Imperative,Need a highly effective learning approach/pedagogy for introducing emerging concepts to undergraduates Concepts: parallelism, algorithms for knowledge discovery, serv
8、ices-oriented design, large-scale data-intensive analytics, design for high-performance and automatic load balancing, large-scale storage and monitoring for access control and performance. Cloud computing encompasses these concepts. Need to provide broad access to these concepts irrespective of the
9、discipline Need to excite and engage the net-generation. The learning should be organic.,6/2/2011,Cloud Futures 2011, Redmond, WA,6,Efficacy of existing methods,Cloud computing discussed as a part of distributed systems course. Ex: I was awarded an NSF CCLI grant for data-intensive computing certifi
10、cate program. But this is just one program, we need a systemic change and a solution that can be adopted and adapted widely Microsoft and other organizations effort during conferences such as SIGCSE to promote emerging technologies Excellent informational overview; however typically not aligned with
11、 the curricular needs and limitations ,6/2/2011,Cloud Futures 2011, Redmond, WA,7,Cloud Models,Amazon EC2: nuts and bolt: good for experienced students (IaaS) Google App Engine: software-based; needs knowledge of services-based programming (SaaS) SalesF; polished applications; more for business-orie
12、nted use (AaaS) Microsoft Azure Platform: (PaaS) Provides a logical framework for modeling and design, and for teaching important computational concepts Offers easy migration into cloud for schools who are already using MS academic licensing (Visual Studio) Provides easy to understand building block
13、s: web role, worker role, blob storage etc.,6/2/2011,Cloud Futures 2011, Redmond, WA,8,Our Approach,We will develop a complete set of self-contained modules customized and aligned to the contents of specific courses. Sustainability through the levels of a curriculum allowing for a spiral mode of lea
14、rning. Educators can use these readily usable modules to introduce Azure cloud computing seamlessly by spending about a weeks lecture (about 3 lecture hours). Modules will be updated with versions of Azure. We will use an application Name2Face as a core-seed application and extend it incrementally t
15、o address issues across curriculum. It is a simple application that is more a metaphor for mapping, correlation and association engine that powers most modern applications.,6/2/2011,Cloud Futures 2011, Redmond, WA,9,Name2Face Details,We have chosen an easy to understand application. Data Structure(s
16、): A list of faces and names. Algorithm: Given a name, the application displays the images of faces associated (tagged) with the name. Demo 1: Here are more Azure details and the demo of the application.,6/2/2011,Cloud Futures 2011, Redmond, WA,10,More details,We use web role, worker role, and blob
17、storage Web role presents the user interface to load the image, tag the image, and to search for an image given a tag/name Web role talks to worker through message passing via a queue Message contains the tags, path to the image Worker role creates a thumb nail of the image (representative of any wo
18、rk) and loads into the blob storage Illustrates: loose coupling, message passing, blob storage service, service authentication through PKI Worker roles can be complex and implement any algorithm being taught in a CSE course.,6/2/2011,Cloud Futures 2011, Redmond, WA,11,Application Architecture for th
19、e demos,Web role,Worker role,Worker role,queue,Blob storage,6/2/2011,Cloud Futures 2011, Redmond, WA,12,Table,Transformative Pedagogy,Includes a set of modules that provide entry points into cloud computing for educators These are ready to use, self-contained modules aligned to the contents of the c
20、ourses at various levels These modules allow for repeated exposure to fully comprehend the broad utility of the cloud and to gain competence in using cloud computing Server-side enterprise-level application development is a simple extension of monolithic single-tier development Next we present a few
21、 representative modules of this pedagogy Detailed report of this pedagogy has been sent to Microsoft and will be available in a future publication.,6/2/2011,Cloud Futures 2011, Redmond, WA,13,Module 1: Data Structures,CS1and CS2 are the first courses in Computer Science; Azure cloud installation and
22、 the Name2Face are introduced with various data structures and algorithms. Develop and deploy on Azure cloud. In memory data-structures. Some representative outcomes: Students exercises can compare performance of arrays, list, maps, trees etc Emphasis on worker role Development and production enviro
23、nments provide the students with professional practices in software development. This module well suited for K-12 as well as Community Colleges.,6/2/2011,Cloud Futures 2011, Redmond, WA,14,Module 2: Algorithms,This covers the programming language course and the advanced algorithms course (at 200 and
24、 300 levels). Azure clouds Blob storage is introduced for large scale storage for Name2Face. Some representative outcomes: Student exercises can study the various sort, search algorithms. Loose coupling; enterprise level development Web role and worker role can be utilized creatively to illustrate p
25、rogramming language concepts; Lambda expression, first class objects, parallelism constructs,6/2/2011,Cloud Futures 2011, Redmond, WA,15,Module 3: Database and Web services,Distributed systems, database systems with 4XX course levels Persistence models, data models, cloud database and database as a
26、service. Learning outcomes: Flexibility of the Azure Table structure Working with SQL Azure data models REST web services access to storage Loose coupling web and worker roles and the storage Scalability of tables and blobs; load balancing, multiple instances of roles Lesson in realizing security th
27、rough certificates,6/2/2011,Cloud Futures 2011, Redmond, WA,16,Module 4: Applications,STEM application domains: data-intensive computing, digital biology for example; Capstone projects Applications drive the advances in technology Expected outcomes: Introducing Azure cloud to other STEM disciplines
28、Provide entry point for non-CSE students Service bus, access control, caching (for CSE),6/2/2011,Cloud Futures 2011, Redmond, WA,17,Digital Biology,20th century is a century of physical sciences; 21st century is a century of great understanding of life Biological infrastructures Data-intensive compu
29、ting for biological problems: specifically genetics, human genome exploration, analysis of high-throughput sequence data, etc. We will discuss some background information and demo a biological index/search application on Azure.,6/2/2011,Cloud Futures 2011, Redmond, WA,18,Unstructured data example (E
30、nglish language),Thes booke is on the tbleThes booke is on the tbleThis book is on the table,6/2/2011,Cloud Futures 2011, Redmond, WA,19,Vertebrate Mammalian Genome,Size 247259409 bytes : 247M (one of many files)TATTAAAAGCACATCTGCCCCAAGAGGCCTTCCCAGACTAAACGCCACTTTTCCTCATTTCCCATTCCCTTCTGCGTCAC CCTAACT
31、TTCTTTTCCCCCGCTTTCCAGCCTCACAGCACTTATGTTTATATTTGTAATGTTATTTATTTATATTGATCTC TGTCTCCCCGACCCCTAGACTGTGAGCTCACTGTGGGCAGGGATTGTCACTGTTTACTGTTGTATTGTGCTTTCCCAGGG GCTTCATACAGTGCTTTGCACACAGTAAGCACTTAAATGCGGTTGGATGAATGAATGAATTTACTACAGCACAGTAGCTG AGAACACAGGGAGACACTGTGATGTGAGGAGGGCTGTATCTGACTGTATTGTACCCACCACAGT
32、TCTTGGCATAGTGTGT AGCACACAGCAAGAACTTATATGTCCCAATTATTTTATTATTTTTTATTATCTGGAATTAGCAATCTACCACTTTGCCCT AGATTGTGAGCCCCACCTGTGTCTTGTATCTACCCTAACGTTTAGTACAGTGCTTTGTATACAGTAAGCACCTAAATACC CCAATTATTATGCATTGTGATGTTCTCGGAATGCTACTAGTCATTTGGATTTTTTTCCATGGGATGATTTAATACTTTCA AGCATTATTCATTGATCTTTCTGGGTTCTAACTGGGTGGC
33、AAGTATAATCAGTCTACTTCAAAGATGGAGAGAACATCCT AGTGTCTGGCAAAGAGGTGGCTAGATCTTAAAGGGACAGGGGAACGCTAGCAAAAAGACTCTAACCCCGATGTTATCTCC TGATCTCCACGCCGCCTCTCTTAGAAGGCCTGCACTGTTTCCTTGCACTATTAATAACAAGGCACTAGCAAAGAAGACTC CCCTCGGAAAAGGATCTCTGAACAACTATTTTGAAAGTTATTTCAAAAGAGTTTTCTATAAGCAATGCATTGAAAATATA AATGCACTTTAAACAT
34、TAAAAACTCCATGGGAAAAAAGTTAGAGTATAAACATACATCAAAATATGCCAAAATAGAACAC ACAGTGTTTATCTGAAGATGTAGCTGTGACACCTAAAATAAAGCATATACGCTGAAAGGATGATATTAAACATAATGGCC TGGATTAGGCCCATAAAACCAATCAAGTTGAAAGGGAATGGCAAAATATGTTGTATAATTTTTTTACATTAAAAGCAAGA TTTCTTGAATTTATCACTCCAACAGATAAAACTTTACTTCATCAAAATGACTTGCATAATGCAGTAGAATAAA
35、TTTCATG ACATATGGATTTTATAAAAGAAAGGGAAAAAGAATGTGCATTCACAGGAGACTCATCATACGGAAGACCTATATTTTTTA AGGTTCTGTGGTTCATTTTTAAGAACGAAACTTAATCTTTAACTCGACCCGCTCTTAAAAATTGTGGTTCATGTGTAAGT TAAAAAGATCAGTGGGTTGAGTTTTCTGGCTGGCAGTTAATCATCTGGAAAAGCTCACATTTATCTGGAACAGAGTAAAT,6/2/2011,Cloud Futures 2011, Redmond, WA,20,Refer
36、ence sequence and query sequence,Assume the reference sequence given in the last slide: Below is a six residue peptide for DPZ, a protein which acts as a tumor suppressor. HPV (human papillomavirus) is a virus that binds to DPZ and degrades it, causing oncogenesis. Codon sequence for: Glycine Leucin
37、e Glycine Phenylalanine Asparagine IsoleucineGGT, CTT, GGT, TTT, AAT, ATT Query sequence: GCTCTTGCTTTTAATATT Now add all the possible mutations, insertions, deletions in the reference sequence Diagnose using (data-intensive) computational methods, the probability of oncogenesis in the person with a
38、given reference sequence.,6/2/2011,Cloud Futures 2011, Redmond, WA,21,Microsoft Biology Initiative (MBI),Microsoft Biology Foundation (MBF) Microsoft Biology Tools (MBT) Excellent resource with many packages with Apache license. From Microsoft Cambridge Research. Fertile source for applications that
39、 can be staged on Azure. This leads us to demo 2.,6/2/2011,Cloud Futures 2011, Redmond, WA,22,Features and Challenges,Great features: Access to cloud computing through Visual Studio PKI-based authentication for storage/server Table storage is very intuitive and highly flexible Challenges The message
40、 queue between the web and the worker roles is a bottle neck The interaction of this queue under .Nets parallelism features need to be further studied by Azure designers. Internal workflow and internal operation of Azure to be clarified: what can and cannot the web role do? Worker role? Best practic
41、es?,6/2/2011,Cloud Futures 2011, Redmond, WA,23,Summary,Cloud is an exciting and high impact infrastructure MS Azure offers excellent resources for teaching cloud computing at all levels It is imperative that our workforce is instructed on this emerging technology to enable them to create innovative cloud-based computing models and novel research directions to solving urgent problems In this talk we demonstrated a systematic approach that provides entry points and pathways to access Azure cloud computing.,6/2/2011,Cloud Futures 2011, Redmond, WA,24,
