Cisco Blogs

Private Clouds are Real – Internetworking for the Cloud

February 7, 2009 - 2 Comments

Just read a very nice piece my Tom Bittman over at Gartner. I agree with Tom, the term Private Cloud, to mean the cloud computing principles and architectures applied internally to an enterprise to achieve economies of scale, simplified and standardized service offerings, scalable growth, and increased efficiencies is fundamentally real.Cloud computing architectures, whether public or private, or frankly ‘virtually private’ (private cloud extending into public infrastructure with enterprise control and trust established) will need a set of networking systems and architectures. As James Urquhart told me once, “you can move the servers to the cloud, you can move the storage to the cloud, but you still have to connect to the cloud.” Cloud Networking could be about building the networking architectures to support private clouds – LAN switching systems with the universal I/O characteristics to allow wire-once infrastructures, the multi-tenant segmentation to allow Layer-2 and Layer-3 isolation, and the operational management characteristics to enable cost-effective operations of a large-scale homogenized infrastructure.But what about ‘Cloud Internetworking’. Internetworking is the core of Cisco, it’s what we do. We link networks together. To me Cloud Internetworking is about enabling the Inter-Cloud, the federation of cloud computing systems between enterprise and provider and one provider to the next. Workload becomes portable, and the Cloud Internetwork embraces this portability and ensures that the elements of trust and control don’t break or disappear with the advent of mobile workloads. Additionally, the Cloud Internetwork ensures that as workloads move they are still reachable via the most efficient path. Lastly, the Cloud Internetwork ensures that enterprises have choices, providers have markets, and infrastructure interoperates.Internetworking is the first-name of Cisco’s core operating systems- Internetwork Operating System. It’s what we do. dg

In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.


  1. You may want to take a look at a counterpoint from Nick Gall, one of Tom’s colleagues:Private Cloud Computing: The Only Thing Real so Far is the Desire because internal IT providers want to build private clouds that offer the benefits of public clouds, doesn’t mean they can. And just because vendors desire to sell internal IT providers something labeled “”private cloud””, doesn’t mean it is one.””There is a good deal of truth in this statement and vendors would want to tread very carefully when pushing “”private cloud”” as “”just like cloud computing, only better”” – I’ve been known call it “”I can’t believe it’s not cloud”” but the term “”fake cloud”” also comes to mind. The true potential of cloud computing is far greater than any “”private cloud”” can offer.That said, evolving virtualisation is a rewarding endeavour and replicating some of the automation from cloud computing providers can give good returns on investment.Looking at cloud computing as an architecture which needs to be wired up to create a “”hybrid cloud”” is a more sensible approach, pulling together resources from multiple cloud providers and internal systems coherently and securely. I’ve been avoiding the whole private/public battleground using this terminology and have done some pretty interesting PoCs with Cisco kit in this area already.Sam”

  2. Doug:All the recent discussions about Cloud Computing Inter-Cloud Computing etc., beg the question: Is Cloud Computing new?It appears to me the IT industry loves hypes — be they thin-clients, client-server computing, object-oriented databases, ASP, ISP, MSP, xSP, virtualization, Web x.0 (where x = 1, 2, or 3), mashups, cloud computing, social computing — which drive the industry, clients get excited, and the analysts and vendors work each other. The problem with us Americans is that we don’t emphasize or focus on history, recent or otherwise.Cloud Computing is nothing new: Thirty years ago it was delivered by Service Bureaus (remember Tymeshare or CDC’s Cybernet?). You submitted a job with punched cards to a computer that was housed who-knows-where, got the results back, fixed any errors, resubmitted your job, and got the final results after a few runs. This was not Cloud Computing, but Fog Computing — you couldn’t see much beyond the green screen, but got your work done! You had a stateless, dumb terminal, usually a 3270; a computer, usually a mainframe; and a network that you didn’t know what it consisted of. Later on, you probably had a TI Silent 700 terminal with thermal paper to input and print out the results, and not a desktop/laptop with bloated and buggy software, lots of memory, disk space, GUI, and a wide-screen monitor. Sure, you didn’t have Gigabit Ethenets, T1, T2, or T3, but, you got your work done, although it probably took you longer than it would today.So, is Cloud Computing going to take over the world? No, but it will play an increasingly significant role. Many defense contractors, banks, and financial institutions do not believe in Cloud Computing because of data protection, security, and privacy issues.Nicolas Carr in his new book The Big Switch”” essentially expounds the concept of cloud/utility computing (aka SaaS) that Larry Ellison of Oracle and Scott McNealy of Sun Microsystems talked about over ten years ago. However, Mark Benioff of actually delivered the solution. Scott even predicted circa 1996 that pretty soon all appliances in your home would be networked and that your light bulbs or fluorescent lights would beg to be replaced before they die. Today, 13 years later, none of the appliances —toaster, microwave oven, refrigerator, freezer, washer or dryer — in my home have an IP address! As the late, great Arthur C. Clarke said decades ago, and I am paraphrasing it, we tend to overestimate the short-term implications and underestimate the long-term implications of a new technology.New technologies, for the most part, supplement existing ones, and not totally supplant them. We believe Cloud Computing will steal some thunder from the traditional perpetual-license, in-house software model, but will not totally replace it.”