IndicThreads.com had their inaugural conference on Upcoming Technology on Aug 20/21 in the city of Pune, India. As the name says, the goal of this conference is to talk about upcoming technologies and Cloud Computing was the chosen topic this time. Harshad & Sangeeta Oak – the driving forces behind Rightrix and this conference made it clear that this conference is not intended to be a training and the aim is to present the latest happening in the cloud world and get the attendees thinking. BTW, Harshad is also an Oracle ACE Director and a Java Champion. The theme of the conference was easily summarized in "Learn, Discuss, Debate, Argue".
Please provide feedback for the sessions at http://u10.indicthreads.com/feedback.
I gave a session on "Running your Java EE 6 applications in the Cloud" and the slides are available below:
I continue to stick to my philosophy of "Code is king" and showed several code samples & screen snapshots. Several attendees told me afterwards that they liked the session because it was well grounded.
The conference had a single track which gives you the ability to attend all the sessions, and there were several of them. Here are my brief notes from some of the sessions I attended. All the slides from different sessions are given below:
And now some notes …
Cloud computing – making the right choice by Kalpak Shah
- Why cloud ?
- No capital expenditure, only operational expenses
- Reduced IT administration
- Elasticity – Fast & inherent scale up/down
- Commoditization of IT – Only storage, only content delivery, only messaging
- Automation using APIs
- Pay-as-you-go for tools & ecosystem
- Reduced time to market, focus on core competency
- Explained the concept of Iaas, PaaS, SaaS, and then Storage-as-a-Service (yet another SaaS)
- Consideration for IaaS
- Combine Support & Infrastructure – Amazon is cheap for infrastructure, costly for support
- Cores are much slower on Amazon, faster on Rackspace
- Good matrix for comparing IaaS vendors
- Comparing PaaS & SaaS vendors
- Development language & available skill sets
- Ease of deployment & maintenance
- Size of vendor & ecosystem maturity
- Tools, monitoring, connectors, adaptors
- Maturity of API & bindings available
- Does vendor allow private cloud integration ?
- Consideration for IaaS
- Ways to Cloudify
- Public clouds – Large datacenters, provide many services (CPU, storage, CDN, databases, etc), can try it, but comes with security, latency & bandwidth issues
- Private clouds Normal on-premise databcenters with cloud with all usual benefits of elasticity, self-service, pay-asyou-go, programmabilty, offerings from VMWare, IBM, Microsoft, Eucalyptus
- Hybrid clouds SaaS Virtual desktops for training Cloud storage for backup etc
- Interesting private cloud platforms
- Eucalyptus – Open-source IaaS cloud computing platform, compatbile with AWS enabling almost seamless movement with AWS, Extremely flexible and easy to administer
- VMWare vCloud – Take the private cloud image, drag/drop on the public cloud and it just works. Platform is consistent & mature and it just works
- Appistry – Application platform for private, public, and hybrid clouds
- Rackspace – OpenStack compute and OpenStack storage
- VMOps
- Azure
- Cloud services operating system
- Provide services across the entire cloud stack – IaaS, PaaS, SaaS
- Development, service hosting & management
- Integrated with exisitng on -premises environment
- Private cloud integration
- Reliance is using Azure for some of their projects
Architecture Challenges in Cloud Computing by Prabodh Navare
- Design for auto-scaling, high performance, failover, data portability, pay-as-you-go
- Vertical scalability – write algorithms that are efficient
- Horizontal scalability – Have the applications as simple/replicable as possible
- Options for private cloud in open Source – Eucalyptus, Ubuntu enterprise Cloud, Open nebula, Nimbus, redhat
Getting started with jClouds by Vikas Hazrati
- Fog, Deltacloud – Ruby-based multicloud library
- Libcloud – Python-based
- Dasein uses jclouds as their base API
- jClouds easy to start
- Simple interface (Map object)
- OSS
- Runtime portability
- Java & Clojure
- Unit testability across clouds
- High performance because of NIO
- Blobstore (atmos, azure, rackspace, s3)
- Compute (vcloud, ec2, gogrid, ibmdev, rackspace, rimu)
- Not 100% but pragmatic portability, dig in with extensions
- Integration with Apache VFS to see the listing of Blobstore
Preparing data for your cloud by Narinder Kumar
- Advantages of Non-relational DBMS
- Scalability
- Replication / Availability (less feature set but more performant)
- Performance
- Deployment flexibility
- Modeling flexibility
- Disadvantages
- Lack of transactional support
- Data integrity is app’s responsibility
- Data duplication is app dependent
- Eventual consistent
- No standardization
- New technology
- RDBMS & Cloud
- MySQL, Oracle, PostgreSQL, DB2, SQL Server are cloud-capable RBDMS
- Microsoft SQL Services and AWS RDS are cloud native RDBMS
- SQL Azure is built on SQL Server and so very intuitive.
- Non-relational DBMS
- Key value stores – Amazon Dynamo (not for public consumption), S3, Project Voldemort, Redis, Scalaris, MemCacheDB, Tokyo Tyrant
- Document stores – CouchDB, mongoDB, riak, Amazon SimpleDB
- Column stores – Google Column Store, Cassandra, HBase, Hypertable
- Graph stores
Day 2 had interesting sessions on Azure by Janakiram M from Microsoft, EC2 by Simone Burnozzi from Amazon and multi-tenancy by Vikas from Inphina.
The Unconference at the end of Day 1 had some interesting topics like Cloud Standards, how Cloud can help fight massive scale diseases, and what a Java stack needs to provide in the cloud.
Over all, I had a great time, enjoyed some great conversations with Dhananjay Nene, Vikas Hazrati, Narinder Kumar, Rohit Naik, Navin Kabra, Manju, Amarpal Singh, and several others. I hope more attendees can join us for an impromptu social gathering in the evening. Anyway, looking forward to participate in the future Upcoming Technology conferences and others hosted by IndicThreads.
Here are some pics from the event:
 |
 |
 |
 |
 |
 |
And the complete album:
