Hi there Mr Roques.

You raise several interesting points here. First, there is nothing about cloud computing that makes it suitable (or unsuitable) for any size business. This issue here is no different than it would be if services were kept at home: it's the quality of the application that matters most. A good application for an SMB is unlikely to suffer from placing it within the cloud. The real question here is does it make good business sense to do it that way? if you're a small business that finds it inconvenient or inefficient for any reason to manage its accounts receivable or payroll locally, it may make very good sense to purchasethat capability from someone who offers that particular software as a service irrespective of where they may be.

The idea of doing backups and recoveries over the cloud certainly seems appealing because a company specializing in this would have a competitive advantage, and thus presumably would be able to offer the service more cheaply than if the company were to try to do it on its own. By way of example, a small hardware store wants to sell hardware, not get into the IT business.

thus, it seems likely to me that broadly-based general applications such as backup and recovery have much to offer to businesses of all sizes when delivered via the cloud, andf that the decision should be based on what makes good business sense.

Your point about sensitive information is also important. Obviously any time we handoff sensitive information to a third party, particularly where such information might have in implications regarding regulatory compliance,we better be damn sure we trust the integrity of the third-party. But data transmitted electronically is easily encrypted at a relatively small penalty in performance, and backups, archives, and replicated DR data have been sent to remote locations for 30 years.

I think the bottom line here is that although your connection to somebody on the other side of the cloud may be virtual, the business relationship has to be based on trust. If the other side of such a potential partnership wins your trust, it should not make a whole lot of difference where in the world the service they offer is physically located.

Hi there, Ninotchka.  I am glad you could get by that "I vant to be alone" thing and join the conversation.

You have asked a really interesting question.  I think we have to look at the issue of "complex applications" with a bit of granularity however.

If the complexity happens at either end of the cloud connection, that is lsikely to have no measurable impact on how long things take compared to when they happen locally.  Where cloud computing is going to run into problems, given the current level of available technology, is likely to be with applications that rely on a lot of back-and-forth chatter between the user and whatever application is out there on the cloud.  We'vealready seen a lot of this problem in more traditional applications that must now work over the WAN, but whichwere originally designed to work over a LAN.  Things happen sloooowy.

A familiar example of this is with any Microsoft Office application, which typically will slow down to a painful level because it expects a LAN-based connection rather than WAN connectivity. The problem with such applications is that LAN protocols are inherently chatty, and must do a lot of handshaking before the first bit of real information ever gets passed along.  This chattiness is quite tolerable over a LAN (where disctances are relatively short), but because electrons travel at a consistent rate the distances involved with a WAN make all that back and forth crosstalk take an intolerable amount of time. 

Fortunately system backups that use the new generation of backup software can (and should) be automated, and therefore do not require a lot of human interaction.  So, there's no reason at all why  backup and recovery over the cloud should not work just fine.

The key take away from all that is that older applications that never were architected for long-haul traffic are unlikely to deliver significant value across the cloud if lots of human interaction is needed for them to do what they do. 

(I am not going to get into the question of WAN accelerators, applications (usually packaged with hardware) that convert LAN protocols to WAN protocols and which compact the data for more efficient transmission.)

One other interesting point I'd like to add:  despite the fact that you obviously have great computer DNA within your personal system, I'm going to disagree with you on a fundamental issue. 

In these days of ubiquitous Web access the idea of "human downtime", at least when it comes to business, no longer has much validity except insofar as it applies to individual humans.  While you and I may be away from the keyboard, people in other time zones are working away vigorously at theirs.  Such is the impact of the Web on the almost all businesses, even SMBs.  Third shifts, historically when backups and other maintenance took place, for many companies are now only marginally less busy than the other shifts when it comes to customer-facing applications.  as a result, that slack time that IT always relied on in many many cases no longer exists.  

Sic transit gloria tertius. Well, perhaps Caesar never said that.

Is cloud computing designed to fit the small to mid size company needs? and only in some industries as I don't see a company with very sensitive information and applications moving to those architectures anytime soon.

Do you think there would be a time where it prove itself even for big companies?

I also think that this is a natural progression for EMC. As you said, who better to eat its lunch than EMC.

Let me run this one by you, since I happen to know that you have some background in telecomm too.

I can't decide to what extent I believe the cloud predictions. Most of the complex cloud-based apps I've tried have been unbearably sluggish, even on a T1. Possibly connections will get faster (or programming will get sleeker) but there must be natural limits to the speed of both.

How is this not an immediate problem for cloud-based backup? Or is that something that people just aren't in a hurry about, since it often happens during human downtime?

Love the new blog...

Hi Murugan. I assume you are the good-looking, shorter one in the picture.

My experience with EMC -- about 9 years' worth at this point -- indicates that in just about every case they will make decsions based on how they see the business case. Whichever route, make or buy, makes the most sense, that's how they will go.

Two countervailing issues are often at work in such situations: First, engineering managers always want to build things (but no successful company only listens to its engineers). And second, in this area of technology time-to-market trumps everything.

It may also be helpful to keep in mind that companies the size of EMC are most effectively understood as being several companies co-existing beneath the same corporate umbrella. They all take their lead from the top of course, but how they implement corporate strategies often bears the stamp of the different business unit leaders. This is a wholly good thing for lots of reasons, not the least of which is that it allows strategies to be (relatively) dynamic rather than just being dogma. It also makes working in such places much more interesting.

EMC has many bright people, and not just in management. What doesn't work with one group may really fly with another.

The bottom line from where I sit -- if it makes sound business sense, they'll do it. And "it" might be just about anything.

Hope that helps.

Mike

Did you by any chance get the feeling that EMC might take the route of acquiring other companies in order to jump ahead in developments related to cloud computing?

Or were they more interested in doing the development work themselves?