UW Information Technology

December 29, 2015

__cloud content__ Top Ten UW IT Cloud Rules

Top Ten UW IT Cloud Computing Rules Of Thumb

So you’ve decided to jump into cloud computing: You are tired of dealing with servers, you need the services without hassling with the hardware, you want access to pre-built computing environments that do just what you need… or you need to spin up 3000 cores for thirty (wall-clock) minutes of parallel processing. For any of the above or for other reasons: Welcome to the cloud and here are some guidelines to help you at UW.

  1. Explore the eScience Institute
    • Come by the Data Science Studio to learn what’s available to help you with your research
  2. UW IT is a resource: Answer questions, find solutions, find related projects, find developers.
    • We see cloud computing as not an end in itself but as a means to an end.
    • We want to get computing hassles out of your way so you can focus on getting your work done.
    • What combination of technologies will work for you? Whatever your task we’ll try and help you get there.
  3. Cloud credit grants are available for research and for education
    • Find out how to apply for cloud credits to get started. Both Amazon Web Services and Microsoft Azure would love to help you do this.
  4. UW IT has umbrella cloud agreements that you can join and benefit from
    • Jump in on a UW IT cloud contract.
    • Whether you are going with AWS or Azure (or other cloud providers TBD)…
    • …these negotiated agreements provide for you to join in.
    • They can save money and bring you other value as well.
  5. Explore the UW IT Cloud Computing pages.
    • Frequently Asked Questions
    • Tools and Technologies
    • Case Studies
    • Deep Dives into particular solutions
    • Data: How it works in the cloud, how much it costs, how it is secure
    • What migrating to the cloud looks like
    • These 10 principles
    • Cloud computing at other Universities
  6. Discover online tutorials and answers to technical questions
    • YouTube tutorial videos
    • GitHub repositories to make your code base stable, shareable and reversible
    • Knowledge bases like MSDN and Stack Overflow
  7. Find and follow best practices
    1. Be sure you’re doing your data security and your work flow the right way.
    2. The cloud is secure provided you configure it correctly.
    3. The cloud can save you money provided you configure it correctly.
    4. The cloud provides high-level services; you don’t need to start from the ground up to get things done.
    5. Here are some abbreviated examples
      • Test apps locally before you deploy them to the cloud.
      • Don’t publish access keys as part of your code base
      • Learn how to sign up for redundant storage to safeguard your data.
      • Assume 90% of your project code base already exists as open source and dare the internet to prove you wrong. You’ll probably be right.
      • Plan to learn how to write test code in addition to your research application code. And document everything. These two steps can also save you time.
      • Use free trial accounts to practice using the cloud.
      • Look for ways of not doing extra stuff. If you program in Python: Figure out how to do that on your chosen cloud platform. Don’t build a web site for sharing data; just build an API. And so forth.
      • Find and learn an Integrated Development Environment that helps you write and test your code.
  8. Ask UW IT to review your system architecture
  9. Build modular
    • Data systems can be developed and tested in a modular fashion with clear boundaries (separation of roles).
    • Do not plan to build everything at once; build according to a logical order.
    • Work backwords from ambitious goals around data-driven collaboration
      • Before you can integrate with other researchers, other systems: Build your analytics.
      • Before you build analytics: Build data query methods.
      • Before building data query methods: Enable data and code sharing.
      • Start by setting up your storage.
      • This sequence — storage to sharing to query to analytics to integration — helps you build your base dependencies first.
  10. Budget for cloud computing
    • Once you’ve figured out whether the cloud can work for you: Put cloud computing line items in your proposal budgets.
    • This budget item will not be subject to Indirect costs