Yesterday I attended a free event arranged by Skills Matter. I was not following the cloud computing revolution that much as it was not happening in Microsoft’s world at a similar level but the rest of the world was more interested in it. Nothing unusual keeping the history of MS in mind. Amazon was in this business from 2002. Skills Matter arranges such free events in the London area and it was an interesting find for me. Cloud computing is defined on wikipedia as:
Cloud computing is Internet ("cloud") based development and use of computer technology. It is a style of computing in which typically real-time scalable resources are provided “as a service” over the Internet to users who need not have knowledge of, expertise in, or control over the technology infrastructure ("in the cloud") that supports them.
The concept incorporates software as a service (SaaS), Web 2.0 and other recent, well-known technology trends, in which the common theme is reliance on the Internet for satisfying the computing needs of the users.
Cohesiveftdelivered this session showing off their virtual servers in the cloud model. Basically they are providing Infrastructure as service. You can think of it as a cheap hotel for applications. If you need to run your application, you can go to their site, configure your own server with your required configuration and software libraries and they will generate a server on the fly for you! Basically all this magic has come about thanks to virtualization technologies which allow you to create software servers independent of the hardware infrastructure running them. These VMs can be scaled and migrated depending on the need.
Their business model is not very appealing to those companies that already own a data center, but to those who would like the processing power without the hassle of maintenance and management, along with the initial investment. This quite simply means that SMEs will approach their services. Cloud has great potential in reducing the cost, increasing utilization and efficiency and simplifying maintenance.
The chief characteristics of the cloud model are:
Customers minimize capital expenditure; this lowers barriers to entry, as infrastructure is owned by the provider and does not need to be purchased for one-time or infrequent intensive computing tasks. Services are typically available to or specifically targeted to retail consumers and small businesses. Device and location independence enable users to access systems regardless of their location or what device they are using, e.g., PC, mobile. Multi-tenancy enables sharing of resources and costs among a large pool of users, allowing for: - Centralization of infrastructure in areas with lower costs (such as real estate, electricity, etc.)
- Peak-load capacity increases (users need not engineer for highest possible load-levels)
- Utilisation and efficiency improvements for systems that are often only 10-20% utilised.
On-demand allocation and de-allocation of CPU, storage and network bandwidth Performance is monitored and consistent, but can suffer from insufficient bandwidth or high network load. Reliability improves through the use of multiple redundant sites, which makes it suitable for business continuity and disaster recovery. Nonetheless, most major cloud computing services have suffered outages and IT and business managers are able to do little when they are affected. Scalability meets changing user demands quickly without users having to engineer for peak loads. Security typically improves due to centralization of data, increased security-focused resources, etc., but raises concerns about loss of control over certain sensitive data. Security is often as good as or better than traditional systems, in part because providers are able to devote shared resources that most customers cannot afford. Providers typically log accesses, but accessing the audit logs themselves can be difficult or impossible. Sustainability comes about through improved resource utilization, more efficient systems, and carbon neutrality. Nonetheless, computers and associated infrastructure are major consumers of energy.
From the characteristics, a few things jump out that point us towards the problems and concerns that are preventing clouds widespread adoption.
Control: When you give up the management of your infrastructure, you also give up control over it. This reduces your burden but also makes you anxious and vulnerable. Now you have to rely on the guarantees provided by the cloud vendor.
Security: This is the biggest concern as data is the most critical asset of a company. They don’t want to put it in any one else’s hand. Great assurances are required here by the service providers, otherwise they aren’t likely to succeed. Not every provider can afford that unless the industry finds a way to simplify this for them.
Dependability: Today a service provider exists, tomorrow they don’t. Companies cannot rely on such services. They need continuity.
Still I believe cloud has great potential and it is possibly the only next method of application development and delivery to be. Hence anyone out there who is interested in development applications for the future should keenly follow this trend as it holds great returns.
For one of the projects I am working on these days, I had to provide image hosting from the database meaning the images will be stored in the the database when the user uploads them to my site and is fetched from the database on request. After looking around for hardly 15mins the solution was there. SQL Server's image type came to the rescue but I read in some article that MS was going to discontinue this type and only keep the varbinary for such purposes. While its available in SQL Server 2005, why not take advantage of it.
Let's say we have the following table structure which represents the different sites of an organization. Each site has its own logo and needs to be displayed on different pages of our site.
Let's first look at how to save the image in the database when it is uploaded from one of the pages in the website (in case its desktop application there is one less step here).
If this were a desktop application I would have probably written:
Let me explain the code here a little. The fuLogoImage is a FileUpload control and the code is simply checking if a file was uploaded through the control when the pages was submitted, using its HasFile property. If it was, then simply get the file as a byte[] and use ADO.Net to write this array to a Image column in the database. In my case here I have used a Typed Dataset, hence the "ta" before the Sites representing the TableAdapter for the Sites table. Notice I simply passed the byte[] without any additional work. You can also simply write a Insert command with the byte[] placed at the right column and ADO.Net will take care of the streaming to database.
Once the file is stored in the database in the LogoImage column, we can then start writing code to display it in our pages. Since the images are being generated dynamically for a given ID, we need a method to bypass the static URL requirement of the image tag. Luckily, there is such a trick and it is called a HTTP Handler. Everything you service in a web server is handled through an HTTP handler, including your aspx pages. If you go to IIS Manager and open the properties of any application, then open the configuration tab, you will see a window like this:
which shows the handler for each type of file the web server (IIS) will be dealing with. Basically, IIS itself does not know how to deal with a file so it passes on the handing responsibility to one of these programs depending on the extension. If I double click the aspx handler description, I see:
So, all http commands for the aspx file are being handled by the asp_isapi.dll extension. This extension is responsible for all the aspx page interpretation, session handling and other features that you like so much. Coming back to our topic, we need to write a handler just like that to generate images dynamically. For that we will use the ASP.Net generic http handler. If you go to your project and add new item, you will see:
Adding this will give you something like this:
The handler implements the IHttpHandler interface which has the capability to handle all the Http commands like Get and Post etc. I you would like session handing capabilities then you can also implement the IRequiresSessionState or IReadOnlySessionState interface. Now, the IHttpHandler contains a function called ProcessRequest where we need to implement our custom logic. I am passing the image ID as a query string parameter to the handler and the code looks like the following:
public void ProcessRequest (HttpContext context)
{
byte[] image;
context.Response.ContentType = "image/jpeg";
decimal siteid = decimal.Parse(context.Request.QueryString["SiteID"]);
if (siteid == -1)
image = GetNoLogo(context);
else
{
ECFormsTableAdapters.Lkp_SitesTableAdapter taSites = new ECFormsTableAdapters.Lkp_SitesTableAdapter();
ECForms.Lkp_SitesDataTable dtSites = taSites.GetDataBySiteID(siteid);
if (dtSites.Count == 0 || dtSites[0].LogoImage == null)
image = GetNoLogo(context);
else
image = dtSites[0].LogoImage;
}
context.Response.BinaryWrite(image);
}Here I fetch the image from the database and then simply write it to the response stream using the proper content type description. In my example I have placed the limit that the image has to be a jpg, as visible from the content type header. The GetNoLogo function basically replaces the image with a default logo that is stored in the file system. The code looks something like this:
private byte[] GetNoLogo(HttpContext context)
{
byte[] imageNoLogo;
string nologopath = context.Server.MapPath("images/nologo.jpg");
lock (lockObject)
{
FileStream fs = new FileStream(nologopath, FileMode.Open,FileAccess.Read);
imageNoLogo = new byte[fs.Length];
fs.Read(imageNoLogo, 0, (int)fs.Length);
fs.Close();
}
return imageNoLogo;
}I have put a lock around the file access to prevent multiple threads from simultaneously opening the file. I should use some caching strategy here but for the moment it is good enough. Now to use it in the asp.net code I define a image tag anywhere in the html
<asp:Image ID="ImgLogo" runat="server" Height="52px" />
and make the src look like this through the code in Page_Load:
ImgLogo.ImageUrl = string.Format("ImageHandler.ashx?SiteID={0}", SelectedSite);and that's it! When the browser renders this image it calls the associated url for image and at the web server this url is a handler that dynamically renders the image using the parameter. Create a site with dynamic images from the database now. Obvious advantages include:
- Single storage of all files.
- Centralized management
- Database searching and meta data storage facility for the images
- Centralized backup and recovery
Some disadvantages include:
- Performance hit
- Single point of failure
But then, the choice is yours. Depending on the situation you have to pick one over another.
This weekend I spent sometime upgrading my site to 1.4 version of BlogEngine.Net from 1.3. Following this post from Al Nyveldt, I upgraded the site first on my laptop and tested for proper working then uploaded to the server. But it was not all hunky dory. Major changes have been made which needed to be adapted to, for example, the new themes support the widget model and all my customizations to previous theme had to be redone to take advantage of the new model. It is not to say that the previous theme was not working, it was, but it was static ofcourse as opposed to the new version. Also one control that I had added to it could not put into the new version because its simply a web control and not made to fit the widget model. I might have to do and upgrade on that as well during the week. Another thing that is giving me unnecessary trouble is that editor for the profile widget. It for some reason always points to the /admin folder and does not allow me to refer images from other folders. To bypass that I had to use the relative path ".." trick to move one level up and point the image to some other location.
One thing that has disappointed me is that there is still the lack of good reporting in the engine. I have used Wordpress before and coming from there I consider this a major shortcoming. There is no reporting on the post visits and because of that no controls for popular posts or regular commenter etc. I hope these features are coming soon. But then, this is also an opportunity for those who need it to build these controls for the community.
I came across this interesting video called "meet the design team" for C# 4.0. After the successful launch of C# 3.0 the design team is already on its way to build the next generation of the language and set the agenda for what is required of it. From the discussion the salient points I can gather are:
Power of dynamic: You know the thing about languages like javascript and VB that you don't have to define everything before hand, like in statically typed languages like C# and its predecessors and you program and define as you go along. Well, the that is the power of dynamic languages. They don't enforce a whole lot of structure on you when writing your programs because they don't care about the program being perfectly typed and thoroughly structured, rather their focus is on the program flow so that you can achieve what you plan to do in the least amount of time. Simple.
Power of functional: If you have worked with any "declarative" language and chances are that you have, then you would know what power functional languages give you. Every developer these days has worked with SQL which is a declarative language and and C# programmers have been exposed to the new LINQ model which is also pretty declarative. The thing about declarative is that instead of writing the program as a series of steps, like a flowchart, where you convert your intent into a well defined structure, you just go ahead and express your intent and say be and it is. Like for example when writing and SQL statement you don't specify "how" it will execute and use indexes and generate temporary tables or loop through a table for every matching row, you just say it should do all that and it does.
Power of concurrency: A point that Anders makes towards the end of the session referring to the Moore's Law, which states that computing power will double every eighteen months, is that the increase of power has taken a shift from more megahertz to more processors because we have kind of reached a limit for how small and fast can we make a processor. So, we are just putting in more of them to meet the demand. Now to effectively adapt to this new trend the languages must have some constructs that allow the programmer to express his "intent" regarding concurrency.
Watch and enjoy.
C# 4.0: Meet the Design Team