Cloud Services with End To End Service Levels

Orange Business Services, Cisco, EMC and VMware have formed an alliance, called Flexible 4 Business. Their plan is to offer four cloud services with tiered end-to-end service-level agreements. 

The services being offered include: 

(1) Private cloud that can be deployed either on customers' sites or on Orange Business Services’ (OBS) network 

(2) data backup 

(3) security-as-a-service and 

(4) unified communications services.  

Good for businesses – I think. 

What attracted us is the phrase “end to end service levels”. 

But there are some immediate questions that come to my mind: 

(1) Are all of the four services fully ready and available in one go? 

(2) Are the SLAs limited to availability and data location? 

(3) How would SLAs vary if the private cloud is deployed on customer’s site and not on OBS? 

These questions need detailed investigation.

Cloud Computing and Corporate Karma

There are so many definitions of the word Karma. And, there are several types of karma – the good, the bad and ... wait a minute ... I haven’t heard about the ‘ugly’ karma.

The phrase Corporate Karma is also not new. There are numerous interpretations of it. There is even a movie by that name I think. To me, majority of the definitions of Corporate Karma seems to have a slightly negative overtone.

For the sake of this article, let’s assume Corporate Karma to be just a metric that tracks current activities and strategies of the enterprise and somehow impacts the future of a business. And let’s confine our focus to Corporate Karma accumulating at IT department.

There are two important questions now:

First how does one accumulate the bad Karma?

Any book on eastern philosophy will answer this question. Overconsumption, accumulation, waste, controlling others, glorifying oneself and not serving others.

Applying this to business, let’s see how a business accumulates karma. It does so, by

-          accumulating assets

-          over-consuming, wasting resources, energy and money

-          establishing control on methods, tools, people and processes

-          creating illusion of mystic aura around nerdy services

-          abusing the word ‘service’

Now, “how does one redeem the bad Karma?” would be our second question.

There are several methods to do so. Not surprisingly, the method that works for you is always unknown, and, seems like a moving target. So, every approach is equally good until you find something that actually works. Hence, it is not surprising to see many people offering different approaches. Even I want slip in one from my side – “Cloud Computing”

How so?

Through virtualisation, you can do more with less and hence won’t need many assets. Through SaaS you can rent software and avoid expensive recurring licenses. Through new pricing models, you will buy only what you actually need. As a result, you could reduce accumulation, spend and wastage.

That’s not all – cloud computing helps further. With the cloud technologies, you could transfer some of the control of your mystic IT to external service providers by aligning your methods and processes with theirs. Thus, you would become unified with the global whole – you are no more an island – your dot is connected with that of others ...

With cloud technologies, you would open up to your enterprise. IT is no more a mystic department hidden behind complex equipment. You become an interesting element of every department. You are no more a nerd or a geek – but just a buyer of services. And you can manage more than an awkward smile.

You will not be busy looking complex or fighting complex code or beastly machines. And slowly the word ‘service’ becomes your mantra.

Oy ... please come down to earth – you seem to be floating! Floating is prohibited. Health and Safety you know.

So, there you go. Cloud computing helps you not to waste resources, not to accumulate assets, to give up control and serve others.

If these won’t help redeeming bad karma, what else would?

Think ... think ... think ...

(Needless to say that this article should be taken lightly and not lightly at the same time!)

AppStores: Future of Enterprise Computing

Smart mobile phones such as Blackberry and Iphone introduced a user friendly approach to browse, select and download a variety of useful – free or otherwise. Now a similar model seems to be on the way for large enterprises, thanks to some recent development in cloud computing especially in the SaaS segment.

What is an Appstore?

In simple terms, an appstore is like a supermarket for applications. You enter one, browse around at intuitively stacked up apps, search for apps or perhaps seek help of online staff to help you pick an application that suits your need.

Just like in a supermarket, you can expect to see promotions of “features” applications, bargains, buyer guides and may be in future, price comparisons. That’s perhaps is a farfetched idea for now – in IT comparison of any two product seems to be similar to comparing apples with oranges.

Just like in the realworld, there is likely to be “buy readymade” or DIY appstores. As their names imply, you may buy what is on offer in the former model and build your own in the latter case. ASDA, M&S are analogous to the former and B&Q, IKEA to the latter.

What types of appstores can we expect?

Several types, actually. But I would say that there would be one appstore within each ‘buying environment’. So, within a large enterprise, there would be a global/regional appstores. Special consortiums of companies such as EADS and mega brands such as Virgin could have their own internal appstores. The Government will have its own appstores – some of them, such as defence, ring fenced for extra security. And, there will be many in the public domain. Major SaaS provider is likely to have one for each industry vertical.

How does an Appstore application work?

There could be a mixture of methods in the way these appstore applications work. Many are likely to run in a cloud based environment – such as a private cloud within the enterprise – or on trusted clouds hosted by SaaS providers.

As the inter-cloud-interfacing APIs and application interoperability matures, there will be applications hosted in one environment or cloud safely and reliably working with applications on other environments or clouds.

It is also possible to imagine meta-appstores which provide just an outer wrapper to other appstores.

Who fills and Appstore and How?

This is an ordinary looking extraordinary question. In the case of large enterprises, procurement department would control the appstore from commercial perspective. There is likely to be a “selection committee” which validates application characters in test environments against arguably tough entry criteria which would involve more non functional characteristics such as reliability, scalability, availability etc.

Vendor maintained appstores will be filled by vendors and their partners.

The process of stacking an appstore in this case is likely to be haphazard – with only a few applications being strategic from vendor perspective.

Enterprise appstores are filled largely based on business requirements.

How to choose applications from appstores?

It is neither an art nor a science. But careful consideration of techno-commercial properties of an application is needed. Certain applications may need ‘subscription’ to a service that runs elsewhere; an application, especially those chosen from a public source may be unreliable and potentially dangerous. Common sense and consultation are the two key things that should guide a user.

DIY Appstores

Some SaaS providers have developed appstores from which you can pick and choose software and service components and build your own situational business application – just like the way you work with Lego pieces. Some service providers like Cordys claim that a simple situational application could be assembled and published in under ten minutes! There would be no need for traditional software design or programming, but thorough testing is always a necessity.

Such applications could help meeting dynamic business requirements and promote agility, innovation at unprecedented low costs.

On the flipside, it could lead to proliferation of uncontrolled business applications over time within the enterprise. Staff turnover, poor configuration management and knowledge management practices will only add to the problem. To remedy this situation, the enterprise must strengthen application reuse and knowledge management practices.

In summary, appstores being in a new dimension into enterprise computing. They help providing an intuitive, controlled, on demand provisioning of IT. Customisable, componentised, DIY appstores bring agile, code free, DIY computing to the enterprise at a risk of application proliferation. They don’t simplify the enterprise IT architecture just yet. But they provide an interesting twist to IT delivery.

Cloud Homework: Checklist For Successful Migration into the Cloud

Having heard so much about cloud computing, I am sure that many businesses would like to taste the cloud for real, even as a pilot project. Where do they start? Is there a “checklist” to follow? There could be many, but I haven’t seen a comprehensive one. Perhaps they are hidden behind many other articles or unpublished or not given much publicity they deserve or I need to read more. Anyways, I thought of writing one.

I don’t want to build a “management checklist” for getting onto the cloud. You know, the typical one which says “Set goals – identify parameters – build a decision tree consider risks, pros and cons – taken an informed decision”. But, I want to build a checklist for a business which has already taken a management decision to go for cloud computing or try cloud computing in a small way.

Don’t forget to perform risk analysis after each step to determine whether you are falling into a danger zone. Also at every step, you must identify a business process to cover failure at that step.

1. Set Expectations Right: Success or failure of your journey depends on your initial expectations. Few examples of expectation setting are shown below:
1. There could be performance trade-offs and latency issues.
2. IT Managers will have to give up some control on their IT applications and data.
3. Manager’s ability to forecast IT usage cost is proportional to their ability to forecast IT usage in terms of bandwidth, storage use and CPU cycles.
2. Cleanup Internal IT: “Cleaning up your internal IT is the first step” says Gartner. The term “cleanup” could mean different things to different companies. Some possibilities are explored below:
1. It could mean knowing your IT architecture, making a list of applications in “real” use, development work going on (authorised and unauthorised), nature of applications (mission-critical, critical, support applications)
2. Trying horizontal integration or “pooling” of similar applications and IT services across departments, cost centres, geographical regions.
3. Identifying dependencies between applications and minimising them as much as possible.
4. Grouping applications that need similar hardware, platform or software environment.
3. Cleanup Data: In this step you need assess the state of your data and condition it for movement into the cloud.
1. Prepare to send the data into the cloud in several iterations/increments/waves.
2. Package the data in such a way that it is independent of underlying applications or data environments. This not only helps data management but also switching data service providers when needed.
3. Ensure consistency in definition of data elements
4. Identify different types of data: data under regulation, vulnerable data, private data and so on.
5. Ensure strong metadata descriptions. This helps data archiving, retrieval and estimating impact of data loss.
4. Attempt Virtualisation: Virtualise as much as possible. Move applications that require similar processing environments to the same server.  Example: Exchange and Mail services. Also ensure that business critical applications are served by a pool of servers, just in case of server failure.

Remember following important points about virtualisation

1. Not all servers, services are ideal for virtualisation
2. You cannot and should not Virtualise everything
3. Virtualisation creates single point of failure!
4. Security challenges remain the same after virtualisation
5. Licenses typically apply to physical servers not virtual ones. So paperwork needs to be checked

5. Distribute Right: Decide what to put where. It is unlikely that a single cloud set-up or a cloud service provider will meet all of your requirements. Hence you will have to live with multiple clouds of different types. Follow a general rule of thumb to decide what application goes where. Here are few suggestions
1. Business Critical Application - Don’t take them anywhere, they stay put
2. Internal Development Environment, if any - Private Cloud
3. Shared Services - Shared Data Grid Services
4. Internet based search – Public Cloud
5. Mail, Collaboration – Consider massively scalable services
6. Consider Readymade Solutions First: Look at existing solutions in the cloud. Is there anything that you can pick up right away that meets your needs? For example
1. Storage for new projects from an IaaS provider?
2. Development environment for new products from a PaaS provider?
3. Do any of the SaaS applications meet an outstanding need of the organisation?
7. Choose your cloud service provider: You need to work with multiple partners to realise your cloud dream. You could use several methods to choose your vendors and partners. Few are examined below.
1. Look at companies that provide technology and associated services.
2. Choose companies based on their experience and maturity. It’s a no brainer. But do remember that there could be new entrants who are more agile and innovative than old biggies.

Remember the following while choosing vendors:

a.    You need to consider yourself to be a team captain in this multi-vendor-partner game of clouds. You need to drive agreements, commitments and own the results.

b.    Some vendors may quote compliance with cloud standards. But cloud standards are still emerging.

c.    Don’t dismiss vendors who were earlier seen as System Integrators, just because system integration for/on the cloud is still nascent. They could help you cleanup your internal IT.

8. Cleanup Processes: Cloud solution needs to be built around your business process not vice versa. Hence reviewing your business processes is quite important.
1. Set up a command protocol or governance model clearly identifying who instructs the cloud service provider and who’ll control service parameters.
2. Review or define new processes that govern commercial aspects of the IT. For example: How does your accounting system need to change in order to recognise value of IT assets?
3. Agree with your cloud service provider, an unambiguous method to track key service characteristics such as availability, performance, security, latency, usage and billing.
9. Clearly Communicate: Concerned stakeholders must be informed and educated. Review this checklist with them and modify it based on their feedback. Be prepared to face resistance, fear and confusion as many stakeholders do not yet understand the cloud.
10. Prepare To-Be image of your IT: If you consolidating all of your cloud homework, you should be in a position to draw post-cloud-migration view of your business/enterprise. If you cannot do so, perhaps it is a good idea to repeat these steps as you may not be ready for the cloud migration yet. If you are able to visualise the post-cloud conditions, then you are ready to test the waters.

It is likely that you will reap some benefits from this checklist even if you decide not to migrate to the cloud. Potential benefits include cleaned up internal IT, data and processes.

And don’t forget to share your experience within the industry.