This week I got an message pointing out that our Egyptian colleagues are unable to work due to the fact they have no more access to the Internet. Indeed, they are unable to access their E-mail and our companies systems that are located in the United States. Indeed, Egypt has cut off most internet and cell service, in an effort to limit the spread of anti-government protest.

My objective with this blog post is not to comment on what happens in Egypt, but look at the impact of such moves to the enterprises working in the country. In today’s world where everything is connected, such move stops the economy in the same way than the truck strike in France a couple years ago. Actually it is even more effective as repercussions are felt even faster.

IT departments increasingly use the internet as the backbone network of their operations. They move applications to consolidated data centres, they limit what employees can do off-line on their mobile devices in an effort to allow platform choice by employees. Aren’t they putting the company at risk?

Internet filtering and ban is nothing new. Researching for this entry, I found articles back to 2006 on the same subject. It puts the future of global politics in question. But I’m astonished that, despite this, companies increasingly rely on that same internet to be the backbone of their communication and their operations. And one finds little about the question on the subject. No pushback, commenting or discussions on the subject. Are enterprises ostriches? Do they put their head in the sand? Do they believe it will not happen within the geographical areas they are present in?

That could be true from a governmental/political perspective, but what about cyber terrorism? We saw some glimpses of that during the latest Wikileaks campaign.

The internet is a horizontal, democratic environment where everything is possible. There are really no rules. It might make sense to put some structure in place, to establish redundant communication links, to set-up a central governance body. But that would require world politics to come up with an agreement, which, as we all know, is rather remote. This leaves companies having to establish recovery scenarios in case the unthinkable happens. And they better start quickly to look at what the alternatives are.

As many of you know, my background is in the supply chain space, and despite the fact I’ve spend the last six months focusing on Cloud, I still have an interest for the supply chain. I’ve had the opportunity to talk to a number of service providers lately and this got me thinking about the parallelisms between cloud and the supply chain.

At the highest level, let me make some comparisons:

• What goods are transported in the cloud? Data
• Where are the inventories? Data storage areas
• What are the transportation routes? The network segments
• What is the transportation capacity? Network bandwidth
• What is the production capacity? Datacenter capacity
• What is the machine set-up time? Resource provisioning time

and I could go on like this. I assume you get the message. And this being said, we can learn from the supply chain lessons to improve the working of the cloud. So, what should we take into account:

1. It’s important to view a supply chain functioning end-to-end, ensuring we are looking at the whole and not sub-optimizing a small portion of the network. Similarly, in the cloud it is important to look at end-to-end services and not limit things to the datacenter under the assumption that the internet functions fine and is outside our control. Hence the importance of looking at both the data center and the network. It is the only way quality of experience can be ensured, in the same way that an end-to-end view of the supply chain is needed to ensure SLA’s are met.
2. Where should inventory be stored in the network. Numerous approaches have been developed to establish locations and sizes of buffer stocks. Similarly, the location of data should be in direct link with the demand. For example, often requested video material should be duplicated in multiple data centers close to the points of consumption, while rarely required ones may be stored in a more central place. The bandwidth required for the long-distance transport can be found more easily as the requests are infrequent.
3. Capacity may be needed on the spot, but often workloads are projected ahead of time. Production and transport scheduling can be used in the same way as it is done in the supply chain. Similar models should be used to identify what can be “produced” when. The unsolicited nature of some of the workloads requires a portion of the capacity to be kept available to respond to this unplanned demand. This is similar, once again, to what is done in some industries, particularly the ones mixing build-to-order and build-to-stock.
4. Provisioning of cloud environments may take time, depending on the strategy used. Here again, things are very similar to machine set-up times for production batches. Depending on the batch type, and in the case of manufacturing, the order in which they take place, more or less set-up time needs to be accounted for. The order in which workloads are executed does not matter at the moment in the cloud as the environment is put back into a common initial state. But as security concerns force more in-depth cleaning of the environment, particularly the storage part, one could envisage a differentiated approach depending on the workloads and their owners. For example, if two workloads of the same owner follow each other, lighter cleaning is required then if it is from competitors.

I could probably go on like this, but would end up boring you. So, let me stop here and conclude. Although supply chains and the cloud are very different things, one can learn from the other. Let me know if you agree with my analysis?

Let me start with a disclaimer. In no way do I support Wikileaks, and the objective of this blog entry is NOT to judge or talk about Wikileaks. However, this beind said, I have seen a couple actions taking place over the last couple weeks that, in my mind, put the functioning of the Internet in danger. That’s what I want to talk about.

The First Amendment of the United States sais: ”Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.” In a nutshell, it guarantees freedom of speech. Now, you can argue that the documents shared by Wikileaks are stolen, so that releasing them is not part of freedom of speech. OK, you have a point, but shouldn’t we then go after the people who did the stealing rather than after the ones who publish?

Secondly, because of the controversial nature of the documents, Amazon shuts down Wikileaks account, becoming a judge of the relevance of the material. Obviously, in a slick marketing message Amazon explains its move by arguing its terms of service. Would be interesting to know since when those terms were in there. Amazon judges that it is not credible that the extraordinary volume of 250,000 classified documents that WikiLeaks is publishing could have been carefully redacted in such a way as to ensure that they weren’t putting innocent people in jeopardy. Again this is a judgement. In my mind that is what courts are for and I do not believe individuals or companies should replace courts.

Beyond this particular fact, Amazon allows itself the rights to discontinue the service of any company that in Amazon’s feeling publishes content that causes injury to any person or entity. Will Amazon discontinue the service of any company that sells weapons or accessories, publishes riffle manuals etc. By the way, they sell weapon accessories themselves as this riffle scope shows. But obviously, that will never put anybody in danger, isn’t it?

The second issue is related to the blocking of the wikileaks.org site name by the US based domain name server (everydns.net). I have the same arguments here. To my knowledge, Wikileaks has not been condemned as a terrorist organization, so why should it suddenly be blocked without any justice intervention? Can the US unilaterally decide who is allowed to be visible on the internet and who not? By the way at least the mirror site wikileaks.ch followed suit quickly.

Beyond this specific case, there is a fundamental issue that appears. As Knowledge@Wharton writes”, “The WikiLeaks Battle: Should Information be Shared or Censored?”, the real question is about what data to share on the internet. This is the first clash between internet freeflow and civil responsibility.  The Internet starts from a concept of trust. It was developed by universities to facilitate collaboration. Today it is way beyond what it was developed for and there is an urgent need to rethink the fundamental principles of the net. Data can easily be routed elsewhere, as shown with the Chinese example of earlier this year, unilateral decisions to block access can be taken etc. I am not even speaking about the hacking and denial of service attacks. The Internet is no longer very pretty, isn’t it?

Your liberty starts where mine ends, but where is that boundary really?

Coming back from the US last week, I read an intriguing article titled “Chinese Firm Hijacked Data” in the Wall Street Journal. The article describes how, sending erroneous messages, a China Telecom routed 15% of the internet traffic through its servers for 15 minutes last April, including the US armed Services, the US senate and companies like Microsoft. Obviously, China Telecom denied any hijack of internet traffic, while the Global Times reported that Chinese experts were saying the report is having little merit because the majority of data in the world is routed through the US.

Frankly, I don’t understand the last statement and its relevance to the case being discussed. My question is simple, was this purely accidental, was it a warning for the US, or was there something else behind this? Obviously we will probably never know the truth behind this. But on a more general basis it highlights the vulnerabilities of the public internet in its current implementation. The Guardian quotes a threat research analyst as saying the capture “is one of the biggest – if not the biggest hijacks – we have ever seen”.

What would happen to the world if suddenly the internet became unusable? Could there be an internet war? This is something that governments will have to address. But what does that mean for enterprises? How can they secure themselves against IP theft ? We all know that confidential data travels across enterprise networks.

Albeit not relevant in the current situation, the point made by the Chinese officials that the majority of data is routed through the US, raises another question. It means that that data is subject to the US Patriot act, allowing law enforcement agencies to search e-mail communication and other records amongst others.

The internet principles are based on trust of equals with no central entity. In the New York Times article already referred above, Lu Benfu, the director of the Internet development Research Center at the Chinese Academy of Sciences, falsely pretends the US is managing controlling the Web Information Flows. He should know better. What is disturbing is that the rerouting began, according to the Guardian, already referred above, with a smaller ISP, called IDC China, before being passed on to China Telecom. Nobody, except the Chinese engineers, know what happen to that data.

This is just one of the examples of the fact we have no control over how data reaches its destination. This is a real issue for location sensitive information. For example, data submitted to export regulations, may end-up being routed through sensitive countries, resulting in breach of compliance. Internet technologists should find a way to protect data from reaching sensitive geographical areas, by for example including in the package headers accepted routing information, or we will have major incidents in the future. So, we really have to speed-up our level of innovation in this area.

Sorry I am getting mathematical here, but I have been talking to many people about this lately and be astonished most people do not look at things in such a simple way. To build a cloud, being it private or public, you fundamentally need three things, one or more data centres that are virtualized and automated, a network that links those DC’s with the end user, and a device through which the end user accesses the services.

When having cloud discussions, people typically jump into the data centres, start talking about software stacks and DC optimization, I do not dispute the importance of that in any way, but feel that, by only looking at the DC, we sub-optimize the approach.

The typical feedback I get is that we have the internet, so why should we worry about the network. Well, I don’t know about you, but how often have I been starring at a small circle that turns around when I try to look at a video on youtube or any other website?

This infinite bandwidth pipe, called the internet, may not be that infinite after all.

This is why I have been looking at what we really need for an integrated DC/Network approach. Two days ago I received an e-mail from Amazon announcing the immediate availability of cluster GPU’s. With that they hope to appeal to high end gamers and to product design and engineering. But how quickly will the complex images and drawing travel over the internet?

Ideally what you want is an integrated provisioning and management of DC/network capabilities. If you require 25 VM’s for a graphic intensive operation, you want to make sure the VM’s are hosted in a DC that has a high bandwidth/low latency network path to your place. And if for one reason or another that link degrades while you are using it, you would like the appropriate rerouting to take place to ensure your quality of experience.

Indeed, the companies that will be able to manage the experience end-to-end, from the device to the DC and back, that will provide superior cloud services to the user, particularly in environments where large amounts of data need to be exchanged between the device and the DC. All is in the equation Cloud=DC + Network + Device.

The other day I was reading an article about  the smart grid. That reminded me of a aection in Tom Friedman’s latest book Hot, Flat and Crowded where he explains how electric cars can serve as buffers, releasing the excess electricity they stored when demand is high, making money on the way.

This is a really interesting model that could apply to cloud. Most servers are linked to each other through the internet. Let now assume they are considered as a resource pool that is available for public cloud activities. When an enterprise does not need server capacity it is released to the global pool of available resources that are then used by the one that needs it. Obviously along the way, the one who uses pays some money to the one who allows his servers to be used.

This is already done, although without the financial elements, in projects such as SETI (Search for ExtraTerestrial Intelligence), but using grid technology. Can we move this one step further and use it within the frame of cloud computing. In March 2009, the London School of Economics came out with an interesting article, titled “Digital Ecosystems in the Cloud: Towards Community Cloud Computing”. In that article they describe the model on how enterprises link together to share compute power.

Obviously, a number of elements need to be addressed to ensure such environment works and we may not address all of these today. In particular the aspects of security has to be taken into account. How can servers been made available for a “public cloud” when they are part of a data centre that may not like public access? How do we ensure we are compliant while using servers without knowing in which geographies they are?

Is the idea of opening up your compute capacity to the world and making a buck doing it, far fetched? Probably in the current state of affairs. No standards are available yet,making it difficult to share infrastructure, security is not yet where it needs to be, the intelligent provisioning of geographically distributed environments, taking the network bandwidth and latency into account, is not fully baked yet. But it might be the way to go to. Any ideays on the subject?

Today I heard a presentation on Bloodhound SSC, the project to break the 1000 MPH miles per Hour) barrier on land. Who the heck is interested in sitting on a rocket, pushed by a jet engine, driving in the desert at such dazzling speed? I wouldn’t want to do it. However there are two aspects of the project that really interest and intrigue me. First, this project is run as a showcase of engineering, and has as key objective to attract new generations to engineering jobs. The second is that, while being built by a community of enterprises, institutes and universities, the project has no issues in making all its data available to the world. This makes it an ideal candidate for a community cloud approach.

Engineering talent, like supply chain talent, are in short supply at the moment, not just in the UK, where this project is run, but across EMEA and the US. Richard Noble pointed out the last time the engineering community had grown drastically was in the 60’s and early 70’s on the back of the “man on the moon” project. So, in his mind a new engineering venture was required. Living in the UK, it could only be breaking the 1000 MPH barrier. And then make sure schools would know about what he does, would follow the venture and get excited about it. Today, nearly 4000 schools, from nurseries to universities, are following the project, building excitement around the country. But if we want to make sure not just UK students get excited, we need to spread the word and ensure teachers around the globe are understanding what is happening. That’s one reason for this  blog entry.

The other point Richard Noble made was that every such car is completely different and that technologies cannot be re-used from one to the other. This means he has no problems making data public. Working with companies, institutes and universities, sharing information and collaborating is key. Wouldn’t this be a good opportunity to develop a community cloud, accessible by all members of the Bloodhound SSC ecosystem. This would ensure everybody is using the latest data, that updates are available instantaneously and that team members  could share their results and experiences.

As far as I could make out from the short discussion I had with him, this is not currently in place, but I believe it would be an excellent project to demonstrate the power of such concept in a real, high profile, project.

So, good luck to Bloodhound SSC and their attempt to break the speed record. Good luck in rebuilding our engineering communities as we really need more candidates moving forward. May the concept of cloud slowly but surely be integrated in the approaches taken by the team.

Cloud computing is making inroads. Companies are experimenting with it. The most advanced do it through their IT departments and focus on areas such as software development & test environments or high performance computing. In the others, it happens despite IT, ultimately putting the company at risk of no longer being compliant.

Once the experimentation phase is over and companies decide to migrate, evolve or re-create specific business applications in a cloud environment, the co-existence of cloud and non-cloud environments has to be addressed. Ideally this should be done in a way that is transparent to the end-user, as he/she does not care about how the functionality is sourced. This brings me to the concept of a hybrid IT environment where cloud and non-cloud applications co-exist.

This requires two key elements to be addressed. The first one is how to present the “services”, and I put the term in brackets. I use it in the SOA (Service Oriented Architecture) sense of the term. The user should be able to address a service catalogue (an apps store in laymen terms), in which he/she finds all applications/services available, regardless of their types. Well, we have been able to integrate legacy applications with the web world for years using web services technologies. We should use the same here. Obviously, there won’t be automatic configuration and provisioning of the service, but the catalogue can link directly to the application through a web service, integrating non-cloud applications with cloud applications.

The other element, which is a little more tricky. How do I integrate my cloud services and legacy applications. It is not easy and in their quest to evolve to cloud, companies need to understand this. Is this the Achilles heel of Cloud Computing as Loraine Lawson puts it? API’s exist, but they are unique to every environment you deal with. So, it’s important to look at migrating blocks that are as functionally homogenous as possible. That limits the integration requirements and makes them easier to implement. Worst case flat files can be transferred from one environment to another.

The concept of aggregation of the available services for the user, regardless of where the functionality is deployed and what technology platform is used, is a critical component in the migration to a cloud based environment. It facilitates the acceptance by the end user while confronting the IT staff with known issues to address (integration, management of heterogeneous environments etc.)

Whether the enterprise decides to build its own private cloud by evolving their data centre, or whether they look at sourcing external, enterprise grade, IaaS or SaaS services, they will have to cope with a hybrid world for the duration of the transformation. And we know that takes a while.

I have been travelling through the US the last 10 days, talking to people about cloud computing and how it will influence our way of working moving forward. As we put more and more information in the cloud and increasingly depend on the cloud to access our services.

Well, last Friday I checked into my hotel room in Palo Alto California around 4PM, and tried to get on the internet to access my email. Interestingly, my PC kept telling me I only had local access. Calling the front desk, I learned that Palo Alto was out of internet access till 7:30 that evening. Can you believe that, we are in the middle of the silicon valley and don’t have internet access.

Beside the point I could not read my email, it demonstrated the vulnerability of the cloud. The internet becomes the critical bottleneck in the cloud, and frankly, I do not believe this is taken seriously enough. I do not see service providers taking serious steps in ensuring availability of quality services all along. What-ever the reason was of this outage, it sounds unbelievable to me there was no alternative routes.

This morning I got an SMS from my admin, she was unemployed, as her internet provider did not deliver. She could not get online and access the services she needed.

Companies are pushing their employees to work from home, on the route etc., while pushing them to use cloud services. But we rely on service providers to link us to the cloud, and guess what, that is where things fall apart occasionally. And there is no way we can complain about this, who should we complain to?    

Last week I had a lot of discussions around the subject of R&D clouds, and the question came up a number of times as whether it made sense to think about it or whether we should just forget.

Having had the opportunity lately to dig a little deeper in one of the lesser known cloud components, namely the network, I believe there is a unique opportunity for the telco’s to deliver such services to their customers.

Indeed, its all about the network. R&D requires low latency as engineers often manipulate large files and graphical objects, and need those at their fingertips. With Moore’s law and the availability of low servers and storage, the network is really becoming the bottleneck. And to my knowledge, there is no Moore’s law for networking.

As the network is operated and managed by third parties (that’s the first thing most CIO’s have outsourced), companies mostly look at this as a given. They get what they get. What I’m trying to argue is that this is no longer enough.

Technologies exist, but are most often not applied, Enterprises should work with their telco and network providers to optimize the end-to-end delivery of low latency networks and ultimately content to the end-user.

Multiple approaches can be taken and choices depends not only on the quality of the network but also on a number of other considerations that may have to do with redundancy of information, IP protection and others.

In their quest to improve engineering productivity, to facilitate collaboration with ODM’s and contract manufacturers, and to speed-up go to market, companies should look at new approaches for R&D, and cloud is one of them. However, that implies technology can match the requirements of the engineers, and this will force enterprises and teclo’s to work more closely together moving forward.

This is a unique opportunity for telco’s to raise up to the plate. In doing so they will not only secure long term customers and increase their revenues, they will also heighten the barrier for new entrants. If they don’t do it, new entrants will eat their lunch, pushing them into the traditional service corner, making them obsolete as those services migrate and transform themselves. So, what are telco’s waiting for?