Server virtualization in MicroLink

Last week I did a presentation on a Microsoft, HP and Intel seminar about how server virtualization is done in MicroLink. I had the honour of presenting the practical experience of virtualization. For me it was a good opportunity to talk to our specialists and compose a few statements about virtualization in practice. So here is our experience:
- Virtualization is good, the technology is working OK.

- The only reason not to use virtualization is licensing. Disk I/O is no longer an issue, but licensing is. For example you cannot use Oracle CPU licenses (too expensive) or Microsoft OLP licenses in a shared virtualization platform.

- As with all Enterprise IT system there are a number of technical nuances when using a "private cloud" solution. Some drivers might not work with virtualization etc. It is better not to discover all these problems yourself but to use the experience of people who have done it before. This way you can save a lot of time and many downtimes.

- Amazon should lower their data transfer rates, otherwise they will not be cheaper than IT hosters virtualization platforms, like for example our MicroLink-s V-Server.

Will Google start charging for its services?

Have you noticed how much we are using Google's services daily? The search, the maps, Gmail, Youtube, Picasa, the books and tens of others. For years now the services have been free, but have you wondered why is that so? Why does a company offer these really expensive services for free? At least for me this sounds strange. I mean Google is a company whose shares are listed on the stockmarket and doing charity is not something a shareholder expects. It is net profit and dividends that are expected by the investors. And Google is no exception.

So what is their strategy? Are they just wasting money or are they just promoting internet services as a charity? Well, this probably is not their plan. They are probably up to something a bit bigger and more serious.
Firstly they still have this really big positive cashflow from the advertising business. Secondly they have an organization of really talented people. These two give an exceptional possibility to build services that do change the world .... and for what the world will later be happy to pay for. To put this more straightforward - Google invests billions of dollars in a service like the maps: they buy the maps, they photograph all the cities in the world, they buy the best satellite photos, they build a huge server farm and a really good interface for the service and they give it away for free. And then a few years later when everybody has become really addicted to their service and there is no way somebody else is able to build a competing offer they start charging a small premium for the service. Something like a dollar or five per month for example. But by then you have hundreds of millions of users and this dollar per users sums up to a hundred million dollars per month.

To gather your opinion I have configured a poll on my (free Blogspot) blog. What do you think, will Google start charging for its services? If yes then when will we start paying the "Youtube, books and search tax"? :-) Please choose the option you think is the right one on the upper right corner of the page.

To give some background here are Google's numbers for the 4th quarter 2009:
- Revenue $6,67 billion
- Net income $1,67 billion
- Number of employees 19 835

Openly about Amazon's cloud

(This article in Estonian.)

One of the biggest Estonian software houses - Webmedia - hold an event last Friday where they launched their new application management service on Amazon EC2 cloud computing platform. I would like to thank the host of the event Tiit Anmann that he invited us from MicroLink also to the event allthough our hosting service is competing with the Webmedia AM service.

In MicroLink we analyzed Amazon cloud services more thoroughly in Autumn 2009 and it was an interesting project. Maybe the most important thing we discovered was that Amazon EC2 and S3 are not competing with MicroLink's IT services but are quite suitable for being used as a resource for our hosting service. As you know a hosting service consists of the following components:
- Data Center
- Network infrastructure
- Storage and Backup
- Hardware
- System level software (operating system, databases, application servers etc.) plus licenses.
- Work: Installation, migration, support, customer support.

Of these components the datacenter, network, storage, backup, hardware and operating system licenses can be either bought by ourself as in the traditional model or be delivered on Amazon platform. So literally speaking now we have an option between buying hardware from Sun, HP or others or use Amazon. The most important part of our service - the high-level IT specialist and customer support - is needed with both options.

Here are the results of our analysis:

Benefits of Amazon EC2 and S3:

- Servers can be started by the system administrator alone and fast. (No need for delivery and physical installation). The same possibilities are currently with our virtual server clusters.
- The service is developed fast and its level is technically high. The VPN, load-balancing and monitoring possibilities are already technically higher level than those of minor and mid-size hosting companies.
- Historically prices have decreased (fast). For example the introduction of reserved instances and the 50% reduction of Windows authenticated license servers in september.
- If the service is down or there is a security incident it is "world news".
- The service has been working well while we have used it for the last 9 months.
- Amazon seems to react to security problems and fix security holes fast.
- The virtual machine images (AMI-s) are OK. There are a number of them for many different operatingsystems.
- Copying of data to EC2 volume works. EC2 volume does not disappear and can be attached to another virtual server.
- Virtual servers are like ordinary servers - we have not yet discovered any problems with any software that has worked on an ordinary server and has not worked on Amazon.
- The computing capacity can be dynamically increased and decreased fast.
- There are a lot of useful software services like billing, e-shop etc.

Minuses of Amazon:
- The data is situated "in the cloud". There is no personal touch or agreement.
- Sending e-mail from Amazon servers is problematic because due to spammers the whole network is in spam black-lists.
- Firewall configuration is simplistic and based on security groups. Groups can be chosen only during startup of the instance and not while it is working.
- Whole action is under one account. So all the systemadministrators must work with the same password and the whole infrastructure is tied to the same credit card. There are not many possibilities to rights management.
- Automatic termination of servers (to save money) seems risky and not worth the money saved. Tools for that must be purchased from third parties.
- If an instance is terminated it dissapears with the data. This is partly a problem of terminology, but might seem surprising for system administrators new to Amazon. And you can be sure this is not a nice surprise. This is also a threat if the operating system hungs.
- Specialists must learn new skills to use the environment.
- Sometimes networking is a problem. The servers are in Ireland and network latency there is 45ms from Estonia.
- Some analytics think that cloud-computing as a business has not yet proven itself and here might be a threat that Amazon closes the service or highers the prices.
- Backup solution must be done differently.
- It is not possible to perform a National security audit (ISKE in Estonia)
- Microsoft license rent model (SPLA) does not support Amazon currently. This is even worse with Oracle.

Price

- Pricing is based on usage, which is good
- All things cost little, but they cost. So the monthly bill can easily grow big. You might forget your server running or start a data copying process. All this will reflect in your monthly bill.
- Keeping costs under control is something that needs an effort.
- Prices are relatively cheap.

Experiences:
- Network access between two Amazon servers must be opened separately.
- After creating an instance it is not possible to change its security group. You must configure this correctly in the beginning.
- Shutting down a server means its dissappearance. If you have not made a bundle of it - it is dissappeared. The server survives a restart, but when the server crashes you must restore it with the help of Amazon support. So here is a risk.
- As the server is like a physical server it is possible to install monitoring agents on it. If you plan to save money by switching the server off for night-time you must take this into account when configuring the monitoring.

MicroLink is looking for training and consulting partners focusing on service industry!

MicroLink's promoter index service has met quite significant success this year in Estonia! All the major banks, retailers, insurance companies, caffees and many, many other companies are using our service.

Now we are looking for a possibility to expand to foreign markets and bring the easy, useful and smart approach to developing a good service to other countries. For that we have developed a following partnermodel:
- MicroLink provides the technology, IT service and helps to do marketing activities.
- Partner, preferrably a trainer or management consultant, talks to customers, helps to promote the idea.
The revenue is shared:
- Partner gets the fee for training and consulting.
- MicroLink gets the 30 eurocents for each response the customer gets.
This model is most certainly negotiable and we are really flexible in developing it!

If you would be interested in becoming our partner then let me know! If you are a trainer or a consultant and your customers have cut their training budgets then this is one way to offer them something new and really useful. We also have a number of international customers who might become your sales-leads.

The promoter index service interface is currently in English, Russian and Estonian. Feedback however can be asked also in Finnish, Latvian and Lithuanian.

Name-tagged photo galleries - what a nice tool for the (secret) police.

Google Picasa recently launched facial recognition for photos. The software works OK and with a decent effort you can "name-tag" all your photos. The name-tags give you nice organizing and sharing possibilities. So quite an ingenious option! The same kind of technology is available for Apple iPhoto and with add-ons also for Flickr. Probably this feature will soon become a must for all bigger photo galleries.

Now lets think about it a bit. A name-, geo- and date-tagged photo (or a movie) in a gallery is an enormous source of information: who knows who, who has been where and when did this happen. If one would gain access to the Google's, Apple's and Flickr databases and all the tagged galleries one could build up information banks that are unprecedented. And if you happen to be amongst the secret police then why stop there. Next connect the photo databases with information from social networks, public internet and the traditional police, bank, credit-card and other databases. So with a simple "select..." query you might get more information about a person than he himself knows. Like what are his 1970's classmates doing at present or what really happened at the party his teenage daughter attended last Friday.

Most surely the amount of geo- and name-tagged photos is growing exponentially with camera mobile phones becoming cheaper the gallery services becoming better. If you have small children and have visited their kindergarten parties you probably have noticed that practically every parent is taking (name-, geo- and date-tagged) photos and making (name-, geo- and date-tagged movies). The lives of our children will be really well documented.

Google's Privacy Policy sais: "We do not sell, rent or otherwise share your personal information with any third parties except in the limited circumstances described in the Google Privacy Policy, such as when we believe we are required to do so by law." Now surely the police and especially all the anti-terrorist units can require Google to do so by law.

A rather nasty side-effect is that besides "the good" also "the bad" will sooner or later gain access to the data. Through hacking or ... just making offers that one cannot refuse to the developers, testers, sysadmins, back-up administrators working for the web-galleries. The only serious obstacle at the moment protecting the data is the share amount of it.

So what are the conclusions and uses of this new development:
- Firstly, live decently. Be aware that everything you do is essentially public information.
- Secondly, this technology could be used more at spots of crisis. For example, why not build a free and easy to use mobile-gallery service for Pakistan and Afganistan. One might even share out GPS enabled mobile phones for free. It seems that the Taleban members like to take photos and show off their arms. Why not let these photos be automatically geo-, date- and name-tagged. :-)
- Thirdly, learn IT! Integration of IT systems and query building will become more and more important for the police and military each day. At least for the next ten years this will create many IT jobs.

Good people management and family life ...a collision?

How to motivate your team of specialists? This is a question frequently asked by modern managers. Everybody knows that a good salary is only a pre-requisite. In order to motivate your team and get the work done much more is needed. You have to make the work interesting, you have to give people freedom of choice and let them develop their own ideas. Modern management is a craft.

Almost all successful companies and amongst others the technology companies have discovered the importance of motivation years ago and have all started to systematically develop the work-place "atmosphere". This is really good as it has made working fun and people much more enthusiastic. If the people love their work and do it with a passion great things happen! Everybody wins - specialists love their work, the companies get good results, customers get nice products....but there is a serious downside to this.

Few companies and few specialists have realized that there is a zero-sum game going on. The object of the game is the employees (and why not also the employers and owners) time. The employers have become really good at this . The best workplaces have an atmosphere where:
- People love what they do.
- People feel that their work is important and they are important.
- Somebody always listens to you. Either your manager, your collegue or the personel manager. They care!
- Because it is fun and important you give 110%.

Now time is a finite resource - so if the workplace wins more of your time and attention then somebody must lose? Who? This is of course different for each one of us, but consider this:
- Does your partner and family have advisors and consultants on "people and relationship matters"?
- Have your relatives been trained on relationship management?
- Do your partner, friends and relatives get paid to have a good relationship with you ... or are they in turn seriously motivated by their employers to pay more attention to their important and fun job?
- Does your partner organize really fun events where you can really party and express yourself ... or are the events dominated by your children expressing themself and you trying to get them scream and dangerously run around less? :-)

So inevitably it is the family and friends who lose in this game. And I would say that this is the second most important reason why we have so many divorces nowadays.

My suggestions:
Firstly everybody can make the TV to be the ultimate loser. Don't watch it! Spend the time won with your family, friends, work and studies.

Secondly the employers should take a second step in the employee motivation plan - think about the families. If we have been successful in making work seem important and fun we most certainly can make family life important and fun.

Thirdly the employers should put some limit to the workhours. And I do not mean switching off the e-mail system at 17:00, but maybe sometimes ask your people to sleep and not to send e-mails, develop software or configure IT systems at 3 o'clock in the night.

NOTES:
* Thanks to Linnar Viik and his great innovation course where we discussed this issue.

** I think that the most important problem why people divorce is the possibility to buy ready made food from groceries and have machines do most of our housekeeping work. This eliminates the important physical reason - hunger and dirty clothes - to live together.

Why you should not build an internet service and drop out of school

Recently there was an article on Arctic Startup encouraging everybody to start an Internet Service. Being old and a little realistic and a father I think I should cool some of the enthusiasm. :-)

Firstly, don't drop out of school. Never! The higher your education the happier your life - the smarter the people you socialise with and the better you understand about what is happening in around. Don't do it even when you make it as an enterpreneur. Nobody likes an un-educated and ill-literate boss. An internet service is no different from any other traditional business and education is one of the few values that a person has. Giving it up just to make a few bucks would be stupid. Besides if we leave out some exceptions (like Bill Gates) then there are not that many rich people who don't have higher education. At some point you just need the general intellectuality. I have worked as a manager of IT specialists and I claim that there is a clear and visible difference between those who have higher education and those who don't. Not having higher education tends to results in a "ceiling of career and salary".

The other problem with Internet Services is the ease of starting them and their globality. If it is easy for you to start one then it is easy for everybody else also. This results in really harsh competition. There are practically no entrance barriers. Neither know-how, capital, geographical nor IP or licensing. You have nothing protecting your revenue. Actually I don't know if there is any other industry where you get so fierce competition. Even the cab companies and restaurants have more entrance barriers (a car, a drivers licence, geography) than Internet Services. As a result that even the profit margins of the PC industry (that are probably less than 5%) look really tempting compared to Internet Services. If every schoolboy can make a service on weekends with zero investment then you end up running the service for free for the rest of its existence. If on the other hand you want to make usability, features, marketing or a large number of users your advantage then be ready to invest - 10-s of thousands of euros.