Thinking metaphorically, my mind goes back to the movie Lagaan, and the scene where the farmers are waiting for the first monsoon clouds to come and break the scorching heat and drought. Will Cloud IT do the same in this recessionary environment? I think so, but you are the judge. This software-driven revolution is set to leverage commodity infrastructure and change the way in which a major portion of IT is consumed and how a major portion of the IT budget is spent.
First there were the mainframes, followed by client-server, 3-tier web architectures, and now we see the beginning of the next major wave in IT – Cloud Architectures. CIOs in 2013, when making infrastructural choices will comfortably deal with cloud solutions as a major part of the mix. This is not simply, the next shiny “IT” thing. Gartner analysts Mark Raskino, John Mahoney, and Patrick Meehan in a recent subscription report adviced CIOs to “start taking cloud seriously. You will need to start leading your organization safely in this inevitable direction, or risk being sidelined by its progress”. Frank Gens, Senior Vice President and Chief Analyst at IDC, has been quoted as saying that “cloud services are ‘crossing the chasm’ and entering a period of widespread adoption”. Fundamentally this is a software revolution. For the first time software is going to enable compute and storage platforms to be easily sliced, diced, and aggregated, taking standard, commodity hardware platforms and delivering IT value, by ensuring never before levels of scalability, flexibility, and efficiency.
So what is it? As with fast emerging technologies, there are many definitions, but Gartner has a fairly board and simple definition: “A style of Computing where scalable and elastic IT capabilities are provided as a service to multiple customers using Internet technologies”. The service can be delivered over the internet as a public/external service or consumed over the intranet as a private/internal service.
It started with applications delivered over the internet(cloud) 3-4 years ago, then called Software-as-a-service (SaaS). With the popularity of SalesForce, Google Docs, Quickbooks-online and a hundred others this sector is here to stay. These applications in the cloud are also cloud applications. Over the past two years, we have gone lower into the IT stack and Amazon started popularizing compute-cycles-for-rent, via its EC2 (Elastic Compute Cloud) service. Underlying this sort of compute offering there typically is a virtualization framework from VMware or Xen. Companies like Elastra, 3Tera, Cassatt have added value over-and-above the underlying hypervisors and delivered software to enable private cloud-compute deployments. To muddle things up further with acronyms, this sort of infrastructural service, is also called, you guessed it, Infrastructure-as-a-service (IaaS).
Over 2008 Amazon’s S3 (Simple Storage Service) has grabbed attention with its service gigabytes for rent by the month. This is a cloud storage service, a public cloud service because it is delivered over the internet, and sometimes referred to as Storage-as-a-service (StaaS). Rackspace, Mosso are others that offer public cloud services. While the Amazon implementation is all about software leveraging commodity hardware, the customer of a public cloud storage service is really buying storage capacity.
In the latter half of 2008, packaged cloud storage offerings have been announced. This really democratizes the space as now enterprises and service providers can build their own storage clouds, and not just have to rely on renting capacity from clouds custom-built by the likes of Amazon and Google. My company, ParaScale, has announced its software-only solution to enable enterprises to build their own private storage clouds, as well as to enable service providers to their public storage clouds. The other pure play cloud storage offering at this point is EMC’s ATMOS.
In 2009, we anticipate rapid acceleration in the number of initial deployments of cloud compute and cloud storage infrastructure amongst the early adopter segment. Even with all the escalating hype about all things cloud, broader adoption is going to start only in 2010 at the earliest. But this is certainly one of the technology areas that is going to see deepening interest in the face of the recession. As customers look to do more with less, data growth and the need for more storage continues unabated, driving the search for more cost-effective ways to address this need.
In a small private survey of the hundreds of early adopters who have registered for our trials, 60 percent were planning to use both private and public clouds, 27 percent only private, and 13 percent only public. This is by no means an exhaustive survey, and really only polls the early adopters, but just points out how customers are starting to think about cloud storage deployments. Looking through the roster of ParaScale registrants shows broad adoption interest. A common misconception is that cloud storage (and compute) is really only relevant to small startup and Web 2.0 companies. While this was true in early 2008, in 2009 interest is much broader.
As the early adopters start deploying cloud storage, they are going to discover that this is a very effective medium for tier 2 and 3. And in many cases, using commodity Linux hardware cloud compute and storage can start very small. Another misconception is that clouds need to be galactic uber-clouds. For example, a small ParaScale cloud storage deployment can be started up for about $10,000, and if existing Linux hardware is repurposed the incremental capex is even smaller. Thus IT shops can take measured steps into the cloud. In traditional enterprises, private storage clouds will allow data-access via familiar enterprise protocols like NFS, FTP, HTTP. In newer Web2.0 type businesses, where storage is a very material cost of doing business, and business and revenue models are unclear, cloud storage can alleviate the P&L pressure.