Blog Battle: Why public cloud is the future of IT

Today we are trying something different in the blog.  We are posting two articles with differing points of view.  The article below takes the position that Public Cloud is more important than the Private Cloud and should take precedent.  The other article, located here, takes the opposing viewpoint that the Private Cloud is more important should be given priority.  

We hope you will chime in with your comments and opinions in the comments.

Let the Blog Battle begin!

Why public cloud is the future of IT

Why not?  Streamline your infrastructure.  Unless you want to manage an ever-growing pile of applications, servers, and storage, the chances are public cloud offers you a better way.  Hoarding stuff under your roof is so yesterday.

Reason 1 - Cost: The traditional “own” vs. “rent” argument always comes up, but it largely misses the concept of innovation.  The downward spiral is of cyclical binge purchases, struggles to digest all that hardware, and the eventual purge.  Huge capital budgets are wasted on over-priced kit that you’ll outgrow in a couple years anyway. Take your backup storage as an example: how many different tape formats have you bought en masse?  4mm, optical disk, 8mm, DLTI, DLTIII, DLT IV, LTO 3, 4, 5….  Every time a new format came out, you probably blew a ton of cash on new hardware and media.  Don’t even get me started on hallways full of self-proclaimed “redundant” disks.  You’ve been left holding bags full of outdated and possibly unreadable data.  Far more efficient to use operating budgets for cloud services that empower you to provision only what you actually need, on demand when you actually need it. 

 

Reason 2 - Human resources: Meanwhile your staff spends their time on low-value activities like unboxing and racking and wiring and patching and repairing and replacing.  Continuing with the tape storage example, how many man-hours have been spent over the last few decades inventorying and duplicating (again!) and ejecting and off-siting and requesting and waiting for the tape before you can start a restore job?  Is that really how you want your information technology professionals spending their days?  The same class of problem carries into other IT disciplines like apps and servers.

So if you’re going to join us in the 21^st century, why not maximize the efficiency of your infrastructure by adopting public cloud services?  There’s a ton of benefit to gain from virtualizing applications and running them on consolidated infrastructure in a highly automated data center.  Just let someone else spend the capital and manpower to run that cloud data center for you!

-        First, you’re going to need to have a clear picture of what applications you have and how they are connected before making changes like migrating servers into the public cloud. Application-aware network performance management, like Riverbed Cascade, can rapidly build application dependency maps to give you visibility to benchmark performance for multi-tier applications.

-        Next, you’ll need to actually move those applications to your cloud provider’s environment, without disrupting service availability. Global load-balancing capabilities in the Riverbed Stingray Traffic Manager can allow you to rewire applications in the background while maintaining service availability.  Your public cloud service will also benefit from the automation that Stingray provides in increasing application performance and reliability, and support greater server throughput for more efficient usage.  As a cloud appliance, you can easily right-size your Stingray deployment, while taking full advantage of your new infinitely scalable infrastructure.

-        Not least, you can bin the tapes entirely, by backing up to the cloud storage provider of your choice.  Whitewater offers drop-in integration with you current backup software, but encrypts and deduplicates the backup data before storing it safely in the cloud.  You can restore locally for common incidents or from the cloud in a “smoking crater” type disaster.  Win, win.

-        Finally, you’ll need to make sure users can still access applications running in the public cloud… which is probably a greater distance from your end users than local servers and data centers were before. Riverbed’s Steelhead WAN optimization technology minimizes the impact of latency – something that adding bandwidth capacity alone cannot do – so that end-users  anywhere can continue to use applications as if they were local.  As an added benefit, Cloud Steelhead and Steelhead Mobile software can reduce bandwidth consumption by 65-95%, providing efficiency and savings that can be funneled back into the business. 

While most hardware pushers slap on a “private cloud” label to keep you over-buying boxes for your data center, the real efficiency and innovation is coming from the public cloud.  We’ll help you get there!

So who's right?  Which do you think is more important to your organization?  Public or private cloud?  Let us know in the comments.

Comments

[Vishwas Manral]

I think we now hear vendors moving out from a public cloud only model to a hybrid one, because of the cost analysis done (Zynga Z-cloud). The general trend seems to be that we should own the base, and public cloud the spike.

Ofcourse to know the base, we need to know the applications, the trends and how they work. Once we know that we can also optimize private IT infrastructure in ways we cannot the Public one.

Comments?? Twitter: @vmanral

[Nik Rouda]

Vishwas,

Great comments! Of course, hybrid cloud will almost certainly be the predominant model, with organizations placing the right resources in right location for the right reasons.

I'm not sure I'd agree public cloud is only for spikes, a lot of other functions can also be cost effective too. The debate is really around deciding which workloads fit the need for price, performance, and control.

The answer may vary a lot by type: apps, servers/compute, or storage, but there are definitely some very attractive use cases.

By the way, Riverbed Cascade can be an extremely useful tool for discovery of over-/under-utilized resources, help map dependencies, and evaluate candidates for public cloud.

You may wish to read more at http://solutioncenters.cio.com/riverbed_hybrid_cloud/

Thanks for sharing your ideas!

Best,
Nik

[DPR]

I think this article is a bit one-sided. It only points out the down-sides to hosting and controlling applications inside the organization. There is a true loss of functionality for application development environments when the public cloud model is used, primarily because there is no secure and high-speed way to ensure that the public cloud can function like a local datacenter and LAN. Are organizations going to pay for gigabits of bandwidth (to reduce down the 10's of gigabytes of bandwidth to their current datacenter) to co-lo sits and SaS sites to somehow achieve cost savings? WAN optimization is an incremental accelerator, but not a magic wand to act as an enabler to allow a 100% public cloud model. Take the case of a typical GIS system with ETL requirements from other datasources. In-house, you have large SANs, large pipes, and the ability to process that data. Off-site in the "cloud" - with multiple cloud vertical providers (app provider, backup provider, etc) - you are limited not only to your speed to one cloud provider, but you are in fact limited by the speed of all of the cloud providers. The IT department could very easily become coordinators instead of technologists, where the majority of their day is spent dealing with the infinite finger-pointing that will begin between cloud providers. All one need do is look at 3rd party WAN connections, such as those from Level-3, etc. where the last mile is typically a local telco. Good luck getting Verizon to troubleshoot the last mile in accordance with an SLA that you have from your local cloud provider. Any CIO/executive who believes a public-cloud only model is the answer will likely have a limited career at the top; a career that will end quickly when their first major outage occurs that they have no control over and no IT staff to help solve (because all of the IT staff members have been moved in-house to cloud providers or other non-IT positions). Look at Blackboard at many Universities, or Google Apps (Commercial Edition) and the number of outages as an example - in which IT directors have been sacked for outsourcing operations that ran perfectly fine inside the organization. The ultimate goal of the "public cloud" is not technological evolution, nor is it cost savings; the goal is simply to keep CIO and other C-level executives in the business of organizational change in the name of resume building - plain and simple (yes, I know that may reduce the legitimacy of this comment, but dig deep and realize it is true in many cases).

[Nik Rouda]

Hi Dana,

Thanks for the thoughtful response, you make some very good points.

You note that this blog post "is a bit one-sided," but I'd say it's extremely one-sided, but that's by design. We thought we'd have some fun taking extreme positions on the public cloud vs. private IT debate and see what happens. The opposing argument is here:

http://blog.riverbed.com/2011/12/cloud-duel-why-mastering-the-private-cloud-should-be-your-top-priority.html

That said, I think the best model is probably somewhere in between, i.e. a hybrid cloud. Organizations should do a careful analysis of their infrastructure, including applications, servers, storage, and networks, and determine from that which mix of resources are best served by public or private delivery mechanisms. The characteristics for evaluation typically include performance, administration, security, availability, and yes, cost, too.

What Riverbed offers is a way better understand your environment and to tip the balance of the above concerns in your favor, in BOTH public and private clouds -- delivering substantial improvements to most environments. This then gives you the flexibility to choose the right place for the right application for the right reasons.

You may wish to read more about this on our "Hybrid Cloud Performance" microsite:

http://solutioncenters.cio.com/riverbed_hybrid_cloud/

As for your closing thought on C-level executives promoting change just to look good: well, I guess we all want to show we're earning our pay. Part of what's exciting about IT is innovation and the continuous search for a better way. Public, private, and hybrid cloud models offer some really interesting possibilities....

Thanks again for your comments, have a happy new year!

Nik