Monday, July 14, 2008

Trailer In The Cloud

I think that cloud computing is a high watch item for 2008. There will be a lot of movement, announcements, and proclamations made this year. I feel that this is one of the boom areas in technology. This is significant because CIOs won’t be able to look away from the allure of SAAS (Software as a Service) and HaaS (Hardware as a Service) and the ridiculous low cost that it can be done for. Outsourcing to India and China is a good example. It may not bring the total benefits envisioned, but the low cost of outsourcing was just too good to pass up. As business tries to get their heads around virtualization they can’t take their eyes off the move to cloud computing. In some respects this is the next step in virtualization. My core data center in where I work is not in San Diego but at another site. Would I care if that data center wasn’t at my company site at all? If it came with the same or better security, performance and reliability, I wouldn’t.

Let me tell you a fictional little story. It’s called “The trailer in the cloud”.

XYZ company made an offered the CIO at Foo Corp. to roll a trailer up to their main office, connect it up to the company network, allowing Foo Corp. to use it’s compute and storage capability for very little $$$. It’s powerful and wonderful. It has a generator on one end for redundancy, is self contained and the latest in technology. XYZ even supplies the people to work inside the beast when needed. “The trailer” even lets you request more compute power and storage when you think you need it.

The CIO accepted this offer and this corporate CIO is very very happy. He he can touch it, smell it, he can feel the power (and the heat) coming out of it. The CIO shouts and grunts in victory, the only way a CIO can, after he has saved the company money by shedding workers and closing a data-centers while still keeping “control” and capability. Hell, he even provided more power, capability and flexibility to the company. This CIO is on the fast-track to greatness.

The CIO then went back into his office and connect to it. He used it and interacted with it on a daily basis (as would his customers in the company). Users just used their stuff, what did they care. The CIO was a happy man.

In a few weeks time “The trailer” moved to the other end of the parking lot. It was still hooked up to the the corporate net. It was still humming along, processing stuff, doing it’s thing and blowing hot air out the back. The hot air that blew out the back smelled like new electronics, power, victory, and supremacy.

The next week the trailer moved just out of eye site of our CIO. The CIO was worried, got a cold sweat and ran into his office. He connect to it, utilized it’s awesome power and assured himself that his data was still there. He took a deep breath and knew that he was still in control. The value, the power, the majestic wonderfulness was still doing the CIO’s bidding.

The next week the trailer moved again. The CIO didn’t even know because he hadn’t seen it since the last move. The trailer rejoined the rest of it’s family, a large grid of computers, acres of them. It is what is known as a compute farm. It’s located close to a low cost power generating station. This farm also is located at the intersection of multiple networks on the Internet and has lower cost cooling and power backup systems. XYZ company saves money by doing this. The CIO at Foo Corp. is totally unaware. No one has called him, his applications are running just like they always have, maybe even a bit faster the morning after the move. The CIO smiles to himself, “It’s bonus time big guy”.

After a big meeting, the CIO calls up a program to request more compute and storage power from “the trailer”. He needs a lot more capability for a new initiative the company is getting ready to implement. He enters in his needs and is asked to confirm the new configuration and cost. The CIO then presses the submit button on his screen and got something quite unexpected. The next page that loaded told the CIO that his request had been completed. Our CIO was amazed, scratched his head and double checked that he entered in his requirements correctly. He did. The system was running, hadn’t lost any of his data, and, amazingly, ready to provide more compute power and store more data. Available disk was instantly enlarged. “The trailer” was ready and waiting.

As the CIO staired at his blinking cursor, smiling and the awesomeness of “the trailer” he picked up the phone to call his boss and tell him that the added storage and compute needs were ready and waiting for the developers to deploy their application. Oh what a feeling to be ahead of schedule.

What the CIO didn’t know was that “the trailer” nudged one of it’s cousins in the farm and instantly added the needs of the CIO. Essentially, “the trailer” grew. The CIO doesn’t realize it, but he is using “cloud computing”. It’s at the farm, it is the farm. To the CIO, he is using “the trailer” even though he isn’t quite sure where it’s parked. To XYZ company, “the trailer” is just a construct, a placebo, a physical something that the CIO can understand, imagine, and feel comfortable with.

The actual trailer that was initially used by the CIO in our story is now at a new customer site where there is a new CIO that is very excited about the cost savings that she is seeing by closing data centers and shedding workers to the land of unknown. All of the real work and storage that our first CIO uses and consumes is being done at the cloud computing center / farm. He doesn’t know it but he’s sharing infrastructure, storage, and sometimes compute cycles with other companies around the world. Those companies all see their trailer in their mind’s eye, wondering how something so small could possibly meet their needs… but it does :)

This story will be repeated multiple times with trailers all over the world. XYZ is happy, their bank account is bursting with cash. The economies of scale and cost of computing is allowing them to make a wonderful announcement to their customers… They are lowering their billing. Customers can continue to use their “trailer” at a lower cost. Customers not only stay with XYZ, now they want more.

This is the end of the “Trailer in the cloud” but another story is about to be written.

I think we will see another good boom when wireless vendors start to deploy on 700mhz. That story is coming soon.

My next article will be to explore Google App Engine and how that may be the catalyst of huge change for computing services and capabilities that companies take advantage of.

– C –

1 comment:

  1. Cloud Computing has some tremendous hurdles to overcome. Ironically, it has scalability problems. Namely, it works great for individuals, reasonably well for small companies that don't want to invest in IT, and increasingly more problematic the larger the organization gets.

    1) Any company that relies on a remote data center, whether their own or a cloud service, has to invest heavily in bandwidth and redundant connectivity. If the WAN goes down - a backhoe incident - there are workers sitting idle and totally at the mercy of the repair. A local data center, in contrast, can keep chugging along even if the Internet outage is a minor inconvenience.

    2) When the "trailer" goes out of view, you lose the warm-fuzzy assurances that nobody is accessing your servers without your knowledge. Can you guarantee the integrity of your data when you cannot control who your provider employs and who has physical access to the machines it resides on?

    3) Your HaaS provider has been issued a search warrant. Are they more likely to cave than if you had your attorney waiting in your own data center lobby to stop them?

    4) If your provider goes out of business, how soon will you know it, and more importantly, how quickly can you get your data back before it becomes inaccessible?

    Give me time, I could come up with more FUD.

    I don't think its a 'fad', but I don't think it is a no-brainer for everyone either.