by: Nagendra Nyamgondalu, Senior Engineering Manager at IBM India and Brandeis Graduate Professional Studies Master of Software Engineering Alum

It was only a few years back that most IT managers I spoke to would smirk when they heard  the  term  “cloud” in  a  conversation.  They  either  didn’t  believe  that  cloud  computing  would  be  viable  for  their  businesses’  IT  needs  or  were  skeptical  about  the maturity  of  the  technology.  And  rightly  so.  But,  a  lot  has  changed  since  then.  The  technology, tools and services available for businesses considering adoption of a public cloud, setting up their own private cloud or treading the middle path of a hybrid one, has  made  rapid  strides.  Now,  the  same  IT  managers  are  very  focused  on  deploying  workloads and applications on the cloud for cost reduction and improved efficiency.

Businesses  today  have  the  choice  of  consuming  Infrastructure  as  a  service  (IaaS),  Platform as a service (PaaS) and Software as a service (SaaS). As you can imagine, these models map directly to the building blocks of a typical data center. Servers, storage and networks form the infrastructure on top of which, the required platforms are built such as databases, application servers or web servers and tools for design and development. Once the two foundational layers are in place, the applications that provide the actual business value can be run on top. While all three models are indisputable parts of the bigger picture that is Cloud Computing, I have chosen to focus on IaaS here. After all, infrastructure is the first step to a successful IT deployment.

Essentially, IaaS is the ability to control and automate pools of resources, be it compute, storage,  network  or  others  and  provision  it  on-­‐demand.  Delivering  IaaS  requires  technology  that  provides  efficient  and  quick  provisioning,  smart  scheduling  for deployment  of  virtual  machines  and  workloads,  support  for  most  hardware  and  of  course, true scalability. OpenStack is an open source framework founded by Rackspace Hosting  and  NASA  that  takes  a  community  approach  to  make  all  this  possible.  It  was  designed  with  scalability  and  elasticity  as  the  overarching  theme  and  a  share­nothing, distribute-­‐everything approach. This enables OpenStack to be horizontally scalable and asynchronous. Since inception, the community has grown to a formidable number with many  technology  vendors  such  as  IBM,  Cisco,  Intel,  HP  and  others  embracing  it.  The  undoubted advantage that a community-­‐based approach brings, especially to something like IaaS, is the extensive support for a long list of devices and cloud standards. When a new type of storage or a next generation network switch is introduced to the market, the vendors have a lot to gain by contributing support drivers for their offerings to the community. Similar support for proprietary technology has dependencies on customer demand and the competitive dynamics amongst the vendors -­‐ this almost always results in delayed support, if that. While proprietary versus open source is always a debate, the innovation and cost benefits that open alternatives have provided in the recent years, has  clearly  made  CIOs  take  notice.  Support  for  a  variety  of  hypervisors,  Open  APIs,  support  for  object  or  block  storage  and  the  mostly  self-­‐sufficient  management capabilities are some of the common themes I hear on why businesses are increasingly adapting OpenStack. Additionally, the distributed architecture of OpenStack where each component (such as Compute, Network, Storage & Security) runs as a separate process connected  via  a  lightweight  message  broker,  makes  it  easy  for  ISVs  looking  to  build  value-­‐adds  on  top  of  the  stack.  All  the  right  ingredients  for  a  complete  cloud management solution for IaaS.

Most  IT  managers  dream  of  the  day  when  every  request  for  infrastructure  is  satisfied  instantly by the click of a button regardless of the type being requested, workloads run smoothly and fail-­‐over seamlessly when there is a need to, resource usage is constantly optimal  and  adding  additional  hardware  to  the  pool  is  a  smooth  exercise.  Business  managers dream of the day when they have instant access to the infrastructure needed to run their brand new application and once it is up, it stays up. Aaah Utopia.

The good news is it is possible here and now.

Nagendra Nyamgondalu is a Senior Engineering Manager at IBM in India. He is a 2003 graduate from Brandeis University, Graduate Professional Studies’ Master of Software Engineering Program.