It seems inevitable that, within the next few years, most IT professionals will have to deal with cloud computing in some shape or form — judging by the amount of publications and discussions, cloud computing sure seems here to stay.
I’ll avoid discussing the exact impact of cloud computing for now, as well as the question which of those 21 different definitions of ‘cloud computing’ is most accurate.

Today, I’d like to take a look at a cloud computing initiative featuring some of the Sybase products that you use in your daily job, such as ASE, Replication Server and IQ.
Sybase has made available a number of Amazon Machine Images (AMIs) featuring these products. Not having used Amazon cloud products before, I decided to start from scratch and see how much time and trouble it took to get ASE and Replication server running within the Amazon cloud platform. You can read my experiences below.
I’d like to state right here that this column is not intended to be an Ad for Amazon, but I must say that they have made it simple to offload some workload into their cloud. You ought to check it out!

In case you still thought of Amazon as an online book-and-CD shop, it’s useful to know that Amazon has long moved beyond retailing consumer products (though they still do that). In recent years, Amazon has opened a new line of business by positioning itself as a cloud computing platform provider. In essence, Amazon is renting out computing capacity-on-demand on their vast global IT infrastructure.
For a small fee (keep reading for details), Amazon lets you use an AMI and you can do whatever you like in it. An AMI is essentially a virtual machine, hosted on Amazon’s cloud network. The AMI can run Windows, Linux or other operating systems, and can optionally use permanent storage in the Amazon cloud. There’s a lot more you can do with Amazon’s cloud offerings (like building and running your own web shop), but I want to limit this column to the Sybase AMIs.

Turning to the Sybase aspect of all this, check out http://www.sybase.com/cloudcomputing . As you can see there, Sybase has made available AMIs for ASE, Replication Server, IQ, and SQL Anywhere. These AMIs are available to anyone signing up with Amazon’s web services; the AMI lets you use the Sybase product in question against a (small) fee, charged by Amazon. The Sybase products in the AMI’s are all running in developer or evaluation modes, and will not require licenses to play with in the cloud. The licensing team in Sybase has come up with a portability method that your account reps can help enable.
Once you have kicked off an AMI, the Sybase product in the AMI is already installed and configured, so you only need to start the Replication server or ASE server in the AMI, and off you go.

Back to my test-driving experiences…. I kept track of the time it took to go from zero to having a running AMI and logging into the ASE server (and Replication Server) inside the AMI.
In a nutshell, it took me 31 minutes to get through the one-time signup for three required Amazon services (AWS, EC2 and S3 — I’ll explain below) and another 23 minutes to figure out how to start the AMI for ASE 15.0.3; it then took 13 more minutes to figure out how to connect to the AMI, and how to log in to the ASE server running in the AMI.
After having gone through this learning curve, doing the same for the RepServer AMI was easier, and took just 30 minutes to start the AMI and connect to it; 22 minutes of this time was just waiting for Amazon to start the AMI.
Finally, getting ASE-to-ASE replication working between two AMIs took another 27 minutes (details below).

The most complex part in all this was to understand what Amazon services to sign up for, and how to choose from the many options you’re being given in the process.
First, you need to sign up with AWS (Amazon Web Services) which is Amazon’s collection of, well, web services, running in the Amazon cloud. This is free, but with AWS alone, you cannot do very much. Therefore, next you need to sign up with EC2 (Amazon’s Elastic Compute Cloud service); this basically provides the on-demand computing capacity for running an AMI. Note that signing up for EC2 requires you to specify a credit card for Amazon to bill your usage to, though the amount will likely be very small if you don’t do heavy lifting: I spent only very little money indeed.
Lastly, you need to sign up for Amazon’s S3 (Simple Storage Service), which offers cloud-based persistent data storage. S3 is also a paid service, though you may not actually need to use it.

Most of this was pretty straightforward, but I had to follow Amazon’s instructions closely. Fortunately, when clicking through from http://www.sybase.com/cloudcomputing to the details of each AMI, there is a “Getting Started with this AMI” section, but this still left me with many questions. Most of those I could resolve once I found the “Amazon Elastic Compute Cloud Getting Started Guide”. For the remaining questions marks, some head-scratching and an intelligent look around on the Amazon pages was enough to keep moving forward (type of problem: some menu items had been renamed and were now slightly different from what Amazon’s instructions said).

Once the sign-up was complete, finding out how to start the AMIs was relatively easy in comparison. Here are some things I found which may be useful if you’re planning to go down this road yourself:

• To start an AMI, you need to sign in with the AWS Management Console first
• In the AWS Management Console, to start an AMI, you must “launch” an AMI image by hitting the “Launch” button
• There are thousands of AMIs to choose from; to list the Sybase AMIs, type “sybase” in the search box in the AWS Management Console after hitting the “Launch” button (select the “Community AMIs” tab)
• When launching an AMI, it can take a while before the AMI is started; in my case it took between 20 and 30 minutes to get the AMI up and running (I’m imagining quite a bit of provisioning must be happening behind the scenes; and obviously Amazon wants you to keep running your AMI). Make sure you hit the ‘Refresh’ button every now and then to see how things have progressed.
• The AMIs for ASE, IQ and Replication Server are all based on Windows (Windows Server 2003 R2); currently no Linux-based AMIs are available for these products (though there’s a Fedora-based AMI for SQL Anywhere). You get Administrator rights on the AMI, so you can mess around as much as you want inside your AMI sandbox.
• When you shut down the AMI, you’re essentially destroying it, so you’ll loose all information in it unless you’ve stored it somewhere else (like on your own local system or on Amazon’s persistent network storage — that S3 thing I mentioned before). As long as you leave the AMI running, Amazon will charge a few cents for every hour (see below) — you start to see the business model here!
• There are various ways to connect to the AMI once it’s running. The simplest is to open a remote desktop connection to it, just as you would when connecting to the desktop of a normal Windows machine; Amazon assigns a unique IP address to each running AMI. You can also connect to the AMI directly from your own system (e.g. with telnet, ssh, or ‘isql’) but this requires generating an X.509 certificate for the required security — you don’t want anyone else to access your AMI, after all. However, I was unable to test this as Amazon kept throwing an error when I tried generating an X.509 certificate; I did follow the polite request to try again later, but something remained broken up in the clouds.
• The instructions for the Sybase AMIs say that the ASE server (and Replication Server) in the AMI will be auto-started when you launch the AMI. However, I found that this hadn’t worked, and I ended up starting the server myself. The server names are SYBASE (for ASE) and DEMORS (for RepServer); you can use Sybase Central, the Windows Services Manager, or the DOS “net start” command to start the servers.
• The ASE server, as well as the RepServer in their respective AMIs have the default blank password for their ’sa’ logins.
• When starting an AMI, a single Sybase server is installed and ready to run. Nothing stops you from creating additional servers though…you have full rights on the AMI, remember?
• The RepServer AMI only contains a RepServer installation, but no ASE. In order to perform some actual replication, you will therefore need to connect to ASE servers outside the AMI. These could run in another AMI or on your own local systems (in the latter case you’ll need to get that X.509 stuff sorted out first). I guess you could also install the ASE Developer’s Edition for Windows on the RepServer AMI, since you have Administrator rights in the AMI Windows environment (I didn’t actually try this myself).
• I decided to prove to myself that I could replicate between AMIs: I set up replication from a primary ASE database in one AMI to a RepServer in another AMI, back to the replicate ASE database in the first AMI (no X.509 certificates or other security provisions are needed for connecting between AMIs). It took me 27 minutes to get this working; most of which was spent on those classic RepServer problems of troubleshooting the connectivity between all components: the SQL.INI files in the AMIs are by default using only “localhost”, and you need to change this to the actual IP addresses of the AMIs themselves.
  Below is a screenshot of the RepServer AMI in full action. You’ll see two DOS boxes: on the left, it’s showing RepServer running in this same AMI; on the right, an ISQL session is logged into the ASE server in the other AMI, and showing a row being replicated from table primdb..mytab to table repdb..mytab (click for full size) :
  

• Currently, the ASE version in the AMI is 15.0.3 ESD#1; the RepServer version is 15.2.

So what about the costs? The currently available Sybase-related AMIs are free from the perspective of Sybase licensing as they come with either a development license or an evaluation license (obviously, you’re bound to the terms of those licenses). However, when using an AMI, Amazon will charge you something for running the AMI. The whole exercise with ASE and RepServer above has so far cost me 32 cents (that’s US$ cents), 30 cents of which went into running the AMIs (12.5 cents/hour; I used the smallest machine size on offer) and data transfer in/out of the AMI (10 cents/GByte). Had I used additional Amazon storage, they’d charged me for that, too (15 cents/month per GByte). Prices vary slightly depending on whether you’re using the US or European part of the Amazon cloud; also, there are various flavours of small/medium/big capacity AMIs, with corresponding price tags.
Obviously, you need determine whether these costs are attractive enough for you to put the AMIs to use, but the startup costs are clearly very low, and you can always pull the plug at any moment.

For now, if you want to try out what Sybase products feel like in a cloud computing environment, there is no reason not to try these AMIs: it’ll be hard to spend more than a few dollars in Amazon costs if you’re doing some light-weight test driving.
In my case, I gained some useful experience with the look and feel of the Amazon platform, for less than US$ 1 — that’s not even worth filing an expenses claim for.

I guess this won’t be the last thing we’ll see about cloud computing and Sybase’s enterprise data management products. Stay tuned.