Cloud Computing -- The Talk of the Town

[admin]

Cloud Computing seems to be one of the trendiest topics to talk about in IT. A successor topic to the virtualization market overall, Cloud Computing seems to be a coalescence of a variety of IT topics into one integrated trend -- perhaps analogous to the universal string theory physicist always look for (sorry for the scientific reference). It's not clear if its real in the sense of being a defined market with distinct product needs or rather a overarching term that more applies to IT architectures than any single product someone might look to buy.

In my mind, the Internet is what is giving birth to Cloud Computing -- providing the ability to cross connect internal computing and external computing. In many respects, it's not a new topic in that companies have for a long time looked to build IT infrastructures that take into account local and remote computing needs. The difference this time is really several fold:

1. Virtualization makes the implementation of services to support Cloud Computing architectures affordable. This means real and sustaining businesses can emerge that provide External cloud services to an IT department.

2. The Internet provides the needed connectivity to make Cloud nodes linkable. This means enterprises can assemble cloud architectures in more of a plug-and-play model than they every have been before.

Some areas of weakness perhaps:
1. Security -- remains a challenge to the software industry overall. Implementing Clouds essentially means data is distributed, access rights are distributed -- two things that leave potential risk factors higher than some companies would like. We need to improve security at all levels of the computing stack for this issue to reduce below a noise level.

2. Reliability -- perhaps on the cusp of being a pro and a con -- with the right data back up models, failures can be handled by altering plug and play services to those that remain up. Perhaps we have lessons to learn with the utility companies who have had to master, as much as they can, models to reroute traffic to maintain high levels of reliability.

3. Cost -- If Cloud is not cheaper than doing it all yourself, why do it. Nonetheless, the pricing/revenue models have to be rich enough for companies to want to be in the cloud services business. Enterprises lean towards service survivors versus fly-by-night startups. After all, that's why IBM and HP continually get business when you think they would not.

Interesting topics to discuss perhaps:
1. What business opportunities exist around Cloud?
2. Is the current views of Cloud worth hanging on to, or should we just treat it as a buzz term and see what's next?
3. Should certain computing infrastructures be institutionalized (eeech -- under government control), given how much society at all levels is dependent on it for (I think this used to be what people saw in the old "utility computing models" talked about early in the first decade of 2000).

Let's have a go at it ...

[akw]

Let me grab business opportunities.

Security is a big issue with the cloud. How do you keep prying eyes out and keep the important stuff from getting out

Provisioning and managing - How do you get the resources you need when you need them and in the right spot

Reporting - How does one show the value of cloud computing to stakeholders

[pselvana]

Since every problem is a business opportunity waiting to be born, here are some other weaknesses with building on the Cloud.

1. Vendor lock-in - Once your data is up on the cloud, it really isn't your data as much as your data residing in someone else's house. If you want to rent somewhere else, there's migration issues that need to be dealt with. Web 2.0 users are seeing this with some similar scenarios like Flickr which offered Rogers Communication Inc customers free "Pro" accounts. When that offer was rescinded, Flickr held the photos at ransom until you paid out of pocket for a Pro account.

2. Network bandwidth/Neutrality - With much of the computation/storage occurring on the other end of the pipe, some services provided will need fatter pipes and guaranteed QoS to maintain a usable system. Both of those requirements cost money. The bandwidth providers (both wired and cellular) will no doubt enjoy the upcoming transition, as will their stockholders no doubt. Services like OnLive require low latency and large bandwidth (720P compressed video) to allow people to play games via the Cloud.


At the same time, the Cloud model does offer some interesting benefits:
1. Thin client - The return of thin clients allows up front costs to be lowered and, if done right, requires upgrading the back-end as opposed to upgrading each and every individual user facing machine. With no customization capable on the front end, a defective client is removed and a spare one replaces it outright with little to no downtime. Benefits are rolled out as needed with the bottleneck becoming the capability for a user to consume them.

2. Economies of Scale - Self explanatory. Added benefits for services among multiple time zones include free load-balancing as users in one half of the world go to sleep as others awaken.

3. Immense Data - Google is great at letting users feed them with data. Data they can and do turn around and analyze to gain value and insight. Other companies can also benefit from having more data than they can compute provided to them - so long as it is secondary to the function and not primary to their service deliverables. Identifying two files are the same or similar, for instance, can lead to removal of redundancies or soft links to clone.

There seem to be a lot of problems as well as benefits with moving to the Cloud. As a model, it might not be right for all. But, if you can get some sure footing in the Cloud, you might be able to raise your business to new heights.

[LoveGolf]

Data privacy seems to be a prevalent issue across many aspects of IT -- look at the problems RIM is having in off shore areas -- will Cloud providers need to provide 'insider' access to their virtual services to local government to protect against terrorism? What if the access gets into the wrong hands -- easily done I am sure. A nightmare for everyone involved. If there was a big breach, would it push people back to owning their own IT or do we just need to weather the storm for a while until the technology settles down?

I am not sure I understand your point on Immense Data -- is it that the Cloud providers can use the data to derive some 3rd party value?

[pselvana]

[LoveGolf]

I guess i see most Cloud services as black box in nature so they should not have access to the data. This is different that ASP services like email where you can control and optimize appropriately -- to me this is not Cloud, especially as ASP is already in place pre-dating much of what is generally called Cloud (or used to be considered Utility Computing using IBM's term).

Given this site is for entrepreneurs, what do you think are the best business opportunities to explore in Cloud if you were to become a participant?

[pselvana]