Reality Checks for the Billowing Cloud Computing Fantasy

Cloud computing has become the hottest, fluffiest and highest profile meme since the dotcom era. The most successful cloud players include in their ranks a bookseller and an advertising/search behemoth. Software companies are making cloud announcements of all kinds.

We owe most of this repositioning to none other than author Nicholas Carr, who has managed to craft a collection of services, technologies and ideas into the Next Big Thing. Yet between the grand vision and a growing collection of visionaries there are plenty of gaps and confusion.

Those gaps can create opportunities and risks for investors.

The Network isn’t Ready

I was on a sobering Infrastructure 2.0 panel in late May about the state of today’s enterprise networks with executives from Cisco (CSCO), VMware (VMW), F5 (FFIV) and Infoblox (my employer), which was a wake up call about the lack of network preparedness for cloud computing. Network vendors could benefit from increasing enterprise investments in cloud, when they happen. As long as they’re ready.

FIRE Infrastructure 2.0 Panel – Photo copyright © 2009 by Sandy Huffaker Jr

Infoblox’s Ness (blog author) and Kagan, with VMware’s Thiele, F5’s Giesa and Cisco’s Gourlay at the Future in Review Conference May 21, 2009

Like the dotcom enthusiasts of their day, many of the cloud champions absolutely refuse to talk about the challenges of cloud for the enterprise, from the lack of automation in the network (especially around addressing), to the lack of security for the elastic Intercloud. Check out Hoff’s blog on Microsoft’s (MSFT) 291 security controls in its cloud architecture; recent Microsoft cloud security claims come across like just another dumb down checklist game.

This makes virtualization security and network automation strategic to cloud. Virtualization security and networking players who establish leadership in these areas would be well positioned to benefit as cloud takes hold.

Microsoft has no doubt taken cloud security seriously and likely has some of the best architects in the business working on securing its cloud architecture, as do Google (GOOG) and Amazon (AMZN). Within the “sausage factory” there are probably some legitimate innovations. But the cloud vendors are doing a substantial disservice to their own marketing by merely dismissing questions and opportunities for deeper dialogue with foil wrapper/cereal box claims.

So we get: “Our Cloud is Secure*" (note asterisk) pronouncements and absurd marketing by checklist for a complex and demanding subject for which there are few genuine experts worldwide. Here comes: “The Bulletproof Cloud”, “Approved by the International Society of Cloud Security Experts” and “The Unhackable Cloud” as banner headlines. They may fly for SMB but they won’t pass muster in enterprise IT shops who understand the complexities of securing elastic cloud infrastructures.

Of course all of this oversimplification is possible because cloud computing today is an amalgam of services and technologies, some of which were rebranded as cloud and others which have yet to be commercially deployed. Others are heavily subsidized by other successful businesses. That makes a single compelling statement both accurate and inaccurate at the same time. Call it a defensible inaccuracy.

As I mentioned, the FIRE Panel (via Vimeo) dug into the networking issues. These are often glossed over by many of the “cloud” players. At Interop earlier that week I was on a cloud panel about what could go wrong. Issues mentioned include security, transport/scale, lock-in and a host of other gotchas waiting for the unwary. These are rarely discussed as the fascination for miracles in the cloud spreads.

The real cloud winners (at least in the enterprise space) will be the companies who introduce real innovation and articulate it to the market in a way that earns them maximum credibility and momentum. They educate the market and deliver the goods. They will overcome the addressing, security, transport, etc issues and enable the powerful Intercloud.

Those who don’t will watch their potentials vaporize.

Disclosure: Long Cisco, VMware. I’m a senior director at Infoblox.

Comments

[Wisdom vs. ...:]

so the cloud is complicated. thanks for completely wasting my time.

Jun 07 02:43 PM

[Gregory Ness:]

Disappointed:

The issue is much bigger than "cloud complication." Its about risks and significant tech barriers to real enterprise adoption and the lack of preparedness for networks. Lets put it this way: if cloud isn't viable for the enterprise how viable will it be for SMB? Yes it can replace some apps (with thin apps and multi-tennant hosting); but if it isn't economical (the elastic intercloud that can follow the moon) and is really just larger VLANs with more VM density... then it won't live up to the "Big Switch" hype.

That issue may noty be that interesting to you, but its an important consideration for enterprises exploring cloud service (versus software as a service or infrastructure as a service, etc) models. And there is a sizable quantity of traditional servers powered up 24X7 (wasting energy) that could be replaced if the issues are properly addressed. If they're not, then we will go on longer wasting the energy used to power these servers while they're not be usied. According to the panel at FIRE, the amount consumed annually by those servers amounts to the entire amount produced by wind (I think it was about 5% that Gouraly mentioned).

Greg

Jun 07 03:25 PM

[Josh B Thom...:]

i read the article to say: It's not as simple as complicated as no one understands it yet. We can imagine [market] the end result, but cannot scale up to that point without assuming currently unavailable intermediate and prerequisite solutions."

P.S. Saying the browser becomes the O.S. is like saying the screen becomes the O.S. That claim ignores the need for an O.S, and reflects the hot air around the cloud-hype that was described in this article.

Jun 07 07:55 PM

[Gregory Ness:]

Josh:

There are solutions that allow for desktops in the cloud. Yes, there is an OS. But the link between hardware and OS is more tenuous.

You can read the article to say that enterprise networks are not ready for private cloud because of security, management/addressing, throughput and storage issues.

You can also read the article (plus others) to say that cloud can decouple the tight relationship between the OS, hard drive and service... introducing new competition to the preload software business. With virtualization and cloud users will have more choices (both in OS and apps and services).

Thanks,
Greg

Jun 07 08:46 PM

[Jasper M:]

MicroSloth couldn't get security right on Internet Explorer, after 6 tries. Why in the World would anyone expect them to get it right in the 'Brave New World' of cloud computing?
I concede that MSFT is big enough to house Some good ideas. But they have so far excelled at preventing them from reaching their customers.

Jun 07 10:05 PM

[Josh B Thom...:]

"the cloud can decouple the tight relationship between the OS, hard drive and service...introducing new competition to the preload software business. "

If there were no anti-piracy devices, the decoupling would already have taken place with the ease of downloading. The break in the business model you are describing is not unique to the OS and hardware relationship.

Jun 07 10:39 PM

[Lightway:]

Great, HONEST article on the development of Cloud Computing. Current industry commentary on the Cloud seems to amount to nothing but cheerleading and hype, without acknowledging the challenges involved in deploying an actual solution. Glad to see some commentary that has developed in the past 2 months that takes a more cautious, realistic approach.

Personally, I don't see the Cloud as the disruptive, landscape-changing platform it's made out to be. These kinds of disruptive movements are extremely hard to predict.

"The real cloud winners (at least in the enterprise space) will be the companies who introduce real innovation and articulate it to the market in a way that earns them maximum credibility and momentum. They educate the market and deliver the goods. They will overcome the addressing, security, transport, etc issues and enable the powerful Intercloud."

Agree 100%. The Intercloud could possibly that next Big Thing. There is a tremendous amount of computing resources out there, just waiting to be harnesses. Collectively, the amount of processing power available on the Grid is staggering.

Jun 08 03:53 AM

[MarkGillCPA:]

I currently suscribe to a few programs that are delivered via "software as a service". The programs work fine for the most part, except they are always slower than the programs that reside on my PC. To me "software as a service" is more a benefit to the provider (and to the companies selling the technology) than to the end user. How will cloud computing be any different?

Jun 08 11:46 AM