Friday, April 29, 2011

Apple preparing a cloud music service

No surprises, they have to react quickly to the Amazon offer for Cloud Drive and Cloud Player.

It's going to be very interesting to see how they make the offer, because, in principle, the cloud and the streaming model will impact the iPod market. If I do streaming I no longer need the music with me. If I do streaming, the iPod can still be the player, but any other can also do it: Android, BB, PC,... you name it.

If the Apple Cloud can only be played with an Apple device it looses part of the benefits of the cloud storage/streaming model. We will see how it's positioned and prices. Apple has very good design, the devices provide some kind of "status", but this is just a service, we'll not see if you're "wearing" the service, so it'll have to be good. The fashion, design, hype components are not so important if there's no device in the equation.

Everything happens very quickly. The MP3 player replaced the walkman or the CD player as a much better way to carry and hear your music, but it's clear that streaming is an even better model: I no longer carry my music or even a special device to hear it. It's always available in the cloud and any device can play it. Is Spotify the iPod killer so Apple has to create something to compete with Spotify and forget about selling more iPods?

Well, we don't know how it's going to be, but the name can be: iCloud. What a surprise, eh? It looks like Apple has payed more than $4 million for the domain.

I can see everybody running to register any possible domain starting with i. I've already registered iCat.

Some lessons from the masters

The other day Facebook made public some of the strategies they follow in its data centers. Google has published a video with some details about the security measures they have in theirs.

Now tell me: Which place is more secure, your data center or the data centers of these guys that are the best of the best at global level?
Also tell me: Who is able to buy cheapest hardware? You buying hundreds (if you're big) of servers per year or this guys buying thousands? Or, as in the case of Google, not even buying servers but just components to make their own server: the "Google Server". Did you know that Google is the third biggest company in the world in server manufacturing? The point is that they are not selling them, to the relief of some.
Who is having better energy consumption per computation unit? Who is paying less for the space of the data center? Who has less people doing system administration?
Can you create your own operative system to optimize the management of the hardware and avoid patching and other expensive support operations?
Do you have access to the best of the best talent for hardware management?

Are you taking my point? There's no way you can compete. They do it better, they do it cheaper, they know more, they have economies of scale that you only dream about. You can still fight for some years against the new cloud model based on misconceptions of security or for compliance reasons, but at the end economic logic will win and I can't see how you can justify your incredible expensive, inefficient, insecure data center while the cloud can do it for you.
(Photo from Flicker by s_w_ellis)

Thursday, April 28, 2011

Amazon cloud failed the other day... So what?

Yep, the Amazon cloud can fail. And, by the way, in case you didn't notice, planes can crash, nuclear plants can have accidents and, in general, there's no way to have 100% security or it's simply just too expensive.
Does it mean that you'll never take another flight or that we have to stop all the nuclear plants? No it only means that you need to understand the reality of life: shit happens!

The royal family never flight all together in the same flight, it's their way to manage the risk of flying. You need to decide how you are going to manage the risk of cloud adoption, as any other thing that you do.

And managing risk is one of those things that they pay us for, isn't it? So go, manage your risk in the new environment as you were doing in the previous one. And don't think that cloud was a fad or that is going to disappear. It's here to stay as a much more effective model than previous ones for a lot of business cases.

You can see the same point well articulated in this article. Just read the SLA and understand what you're signing for. It's interesting to see that Amazon didn't fail to comply with the SLA at all along the crisis. Likely the customers will ask to see those other services that are not exactly EC2, but that are becoming critical, in the SLA in the future. So it’s a good learning for everybody.

Have you heard those voices saying: we already beware you about cloud computing! We were right, now that Amazon is down you see it. I cannot believe how somebody that pretend to be intelligent can do such a simplistic analysis.

And Amazon cloud will fail again, I bet for it. And, and Azure and all the others either have failed before or will fail in the future. Again and again. Shit happens.

Tuesday, April 26, 2011

Google Docs: Start with your personal cloud

Google has launched many services for particular users over the years: search, gmail, Google Calendar, Google Reader, Google Docs, Google sites, buzz, finance, groups, iGoogle, etc. The list is really impressive and you can use all this for free, but that's not the point of the post.

They have gone one step further in making the adoption of Google Docs much smoother. Now you can edit your documents with MS Office applications and save everything into the Google Docs cloud. It can save only to Google or it can keep a version localy and a copy in the cloud. You can also keep versioning of the documents if you want. You just need to install a plug-in for Office that you can find here

This way, as you work with your documents more and more of them end up in the cloud and available every where you go.
Same tools, same computer, cloud storage. Transparent for the user. Isn't it clever?

"De-perimetirization" is (was) the key

This topic was raised already several years ago by the Jericho Forum, that is a special group dedicated to security inside The Open Group. One of my colleagues in Lilly was a member of the Jericho Forum and he was already raising the issue: The perimeter was no longer a sustainable security model if you need to open your doors to vendors, partners, alliances, employees working from home or from the field,....
The workers don't want to connect using code generators gadgets, VPN and three level authentication, they want to connect as they connect to their banks (at the most) and collaboration with other companies and universities becomes a critical need to survive.

In this situation you can no longer relly on the perimeter to keep the bad guys out. In this model, if you're in, you have access to almost everything. If you're out, you get nothing. Well, it was clear some time ago, but now with the explosion of cloud computing is even more clear that the model can no longer be sustainable, as they say in this article in Computer World.
Security has to be something that is attached to each asset and can not be based in a perimeter that in many companies no longer exists. So this is: De-perimerization, and incredible difficult to pronounce word.
It was clear at least two years ago, what has your company done to change the security model? Do you still have a perimeter? How are you going to keep it if you use cloud computing?

The security guys should be happy, they still have a lot of work to do.

Saturday, April 23, 2011

Thursday, April 14, 2011

New talk about cloud computing

What a surprise, right?

Well, clearly Cloud Computing is the topic with most Hype right now, it's like SOA four or five years ago.

I'll be talking in a Cloud event on the 18th of May. I'll tell you how it goes.

I'm one of the first ones to talk along the day, I don't think that most of the hardware vendors will like what I have to say about private clouds, hybrid clouds and public clouds, so we will see what happens and if I survive.

(Photo from Flicker by Fractal Artist)