Staff can update the at-risk status by clicking on the “Update At Risk” link in the actions list.

The post wisiPlan eILP At Risk Management appeared first on .

IaaS and SaaS eco system is well understood today.   In SaaS, an organization  will maintain the service, typically be subscription based revenue model.  The service might allow potentially some extensibility.  Likes of salesforce take it a bit further and make it into a platform which blurs the lines between SaaS and PaaS (with salesforce.com being both SaaS and PaaS)

[I will use the term CloudOS to nearly mean PaaS in this discussion that follows.  This helps us in thinking about the issues in a way similar to OnPremise OS.]

What is a very interesting case to think through is when Software is sold for the Cloud OS.  What I mean by this is, a software bundle is deliverd by the vendor to be installed by the organization on its Cloud OS subscription or instance.   This is not a very new concept as we do have some bits and pieces of such a thing before, such as forum software being downloaded and installed on a server. 

For the sake of discussion

• Let’s say Wignite is in the business of selling enterprise-networking software for Cloud OS called EnterpriseSocial.
• EnterpriseSocial software is bought by a company name FMG Inc.

Now,  how will this software be delivered, installed, updated, managed and what would the licensing agreements look like.    Following are the some of the interesting points that come up when we think of selling software for a cloud OS.

1. Would the revenue model still be subscription based or will it go closer to traditional upfront software cost and then some AMC costs.
2. Will the licenses restrict how many frontend or backened servers that can be running EnterpriseSocial.
3. Which Cloud OSes will EnterpriseSocial support? Windows Azure, AWS, AppEngine or something from IBM? any new middlewares?
4. What will be the software controls that admins have to allow elastic scaling of servers and other resources that the CloudOS supports?  Will this be able to take the changingCloudOS’ infrastructure costs?
5. Who owns the installation of software updates?  Can Wignite force installation of updates?  will there be a 3rd party cloud update service or will it be owned by the Cloud OS?
6. Since we will be giving out the software for installation and cannot afford to give the source code, should we avoid writing the software in php, ruby (and stick to C#, java or C++ or C!)
7. Since the code is for a CloudOS, will there be better controls to tackle software piracy (such as the cloudOS supporting a way to check with wignite server on validity of software and automatically shut it down)
8. Will EnterpriseSocial co-exist with other software on the cloud servers?  This may seem obvious initially, but remember we are also talking of other things such as auto-scaling being supported by EnterpriseSocial.  Now, if FMG is running 4 such applications on CloudOS sharing some of the cloud servers and if all of them autoscale on their own, all hell will break loose.
9. Will CloudOS appstores take off in a big way?  What control and checks will they have on the software that gets published through the app stores?

 Nobody has definitive answers as the Cloud OS market is still evolving, but I am sure all of us who are working on the cloud will have some views on each of these.

What are your thoughts on the issues I mentioned above?  Any other important issues that I have missed?

The post Selling software instead of SaaS for Cloud OS appeared first on .

One might wonder latency can’t be reduced between two locations using better technology.  Though most of the readers might know, I thought I should explain this for those who didn’t think about this.

Just take the example of trying to access a web page from New Delhi from a server located in San Franciso.  Rough distance between the two cities is 12,348 KMs(from mapcrow).  The maximum speed at which we can send information is limited by speed of light which is c = 3×10^8 meters/second in vaccum.  Our communication is typically over fiber optic cables which is essentially glass.  In this medium, the speed of light is approx 2×10^8 meters / second.  (It is to be remembered that no other means of communication can go beyond speed of light)

Thus, to send information from New Delhi to SF and back, the information can at best travel at 2×10^8 meters / second over a distance of 12,348 kms and back.  This means, a minimum round trip latency of  roughly 123 milli seconds.

Thus, whatever we do, we can’t reduce the round-trip latency beyond 123 milli secs unless we move the server to a location closer to the user.

The post Why latency can’t be reduced beyond one point appeared first on .

Step1: In service definition file(ServiceDefinition.csdef), add the following lines

 <Imports>
 <Import moduleName="Diagnostics" />
 </Imports>
 <Startup>
 <Task commandLine="FixDiag.cmd" executionContext="elevated" taskType="background" />
 </Startup>

Step2: In ServiceConfiguration.cscfg file, add the value to the configuration settings as follows

<ConfigurationSettings>
<Setting name="Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString" value="PASTE YOUR STORAGE CONNECTION STRING" />
</ConfigurationSettings>

Step3: In WebRole.cs, in the OnStart() method, add the following lines

public override bool OnStart()
{
CloudStorageAccount.SetConfigurationSettingPublisher((configName, configSetter) =>
{
configSetter(RoleEnvironment.GetConfigurationSettingValue(configName));
});

// Obtain a reference to the initial default configuration.
string wadConnectionString = "Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString";

CloudStorageAccount storageAccount = CloudStorageAccount.Parse(RoleEnvironment.GetConfigurationSettingValue(wadConnectionString));
RoleInstanceDiagnosticManager roleInstanceDiagnosticManager = storageAccount.CreateRoleInstanceDiagnosticManager(RoleEnvironment.DeploymentId,
RoleEnvironment.CurrentRoleInstance.Role.Name, RoleEnvironment.CurrentRoleInstance.Id);
DiagnosticMonitorConfiguration config = DiagnosticMonitor.GetDefaultInitialConfiguration();
config.ConfigurationChangePollInterval = TimeSpan.FromSeconds(30.0);

//transfer the IIS and IIS Failed Request Logs
config.Directories.ScheduledTransferPeriod = TimeSpan.FromMinutes(1.0);

// Configure trace logs
config.Logs.ScheduledTransferLogLevelFilter = LogLevel.Undefined;
config.Logs.ScheduledTransferPeriod = TimeSpan.FromSeconds(30.0);

//set the configuration for use
roleInstanceDiagnosticManager.SetCurrentConfiguration(config);
}

Step4: Add the following lines to web.config and copy the file web.config  to WaIISHost.exe.config

<system.diagnostics>
<trace>
<listeners>
<add type="Microsoft.WindowsAzure.Diagnostics.DiagnosticMonitorTraceListener, Microsoft.WindowsAzure.Diagnostics, Version=1.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" name="AzureDiagnostics">
<filter type="" />
</add>
</listeners>
</trace>
</system.diagnostics>

Step5:  Add the following two files to your project and set their property “Copy to Output Directory” to “Copy Always”

File1: FixDiag.cmd (contents of file below)

powershell -ExecutionPolicy Unrestricted .\FixDiagFolderAccess.ps1>>C:\output.txt

File2: FixDiagFolderAccess.ps1 (contents of file below)

echo "Output from Powershell script to set permissions for IIS logging."

Add-PSSnapin Microsoft.WindowsAzure.ServiceRuntime

# wait until the azure assembly is available
while (!$?)
{
echo "Failed, retrying after five seconds..."
sleep 5

Add-PSSnapin Microsoft.WindowsAzure.ServiceRuntime
}

echo "Added WA snapin."

# get the DiagnosticStore folder and the root path for it
$localresource = Get-LocalResource "DiagnosticStore"
$folder = $localresource.RootPath

echo "DiagnosticStore path"
$folder

# set the acl's on the FailedReqLogFiles folder to allow full access by anybody.
# can do a little trial & error to change this if you want to.

$acl = Get-Acl $folder

$rule1 = New-Object System.Security.AccessControl.FileSystemAccessRule(
"Administrators", "FullControl", "ContainerInherit, ObjectInherit",
"None", "Allow")
$rule2 = New-Object System.Security.AccessControl.FileSystemAccessRule(
"Everyone", "FullControl", "ContainerInherit, ObjectInherit",
"None", "Allow")

$acl.AddAccessRule($rule1)
$acl.AddAccessRule($rule2)

Set-Acl $folder $acl

mkdir $folder\FailedReqLogFiles\Web
"placeholder" >$folder\FailedReqLogFiles\Web\placeholder.txt

mkdir $folder\Logfiles\Web
"placeholder" >$folder\Logfiles\Web\placeholder.txt

# Get a list of the directories for the regular IIS logs.
# You have to wait until they are actually created,
#   which is why there's a loop here.
# Just keep looking until you find the folder(s).

$dirs = [System.IO.Directory]::GetDirectories($folder + "\\Logfiles\\web\\", "W3SVC*")
$ip = (gwmi Win32_NetworkAdapterConfiguration | ? { $_.IPAddress -ne $null }).ipaddress

$ip
echo "dirs.count"
$dirs.count
while ($dirs.Count -le 0)
{
Sleep 10
$bs = (new-object System.Net.WebClient).DownloadData("http://" + $ip[0])
echo "in the loop"

$dirs = [System.IO.Directory]::GetDirectories($folder + "\\Logfiles\\Web\\", "W3SVC*")
echo "dirs"
$dirs
echo "dirs[0]"
$dirs[0]
echo "dirs.count"
$dirs.count
}

echo "after while loop"
echo "dirs[0]"
$dirs[0]

# Now set the ACLs on the "first" directory you find. (There's only ever one.)

$acl = Get-Acl $dirs[0]
$acl.AddAccessRule($rule1)
$acl.AddAccessRule($rule2)

Set-Acl $dirs[0] $acl

This should set up your diagnostics, diagnostics transfer to the corresponding blob & table storage in azure.  In my Part-2 of this post, I will describe the basics of windows azure diagnostics and where to find what.

Most of the above code is taken from http://robindotnet.wordpress.com/2011/02/16/azure-toolssdk-1-3-and-iis-logging/

The post Setting up diagnostics with Windows Azure SDK 1.3 – Part1 appeared first on .

Full IIS support and consequent changes over 1.2

http://blogs.msdn.com/b/windowsazure/archive/2010/12/02/new-full-iis-capabilities-differences-from-hosted-web-core.aspx

Configuration changes wrt diagnostics

http://convective.wordpress.com/2010/12/01/configuration-changes-to-windows-azure-diagnostics-in-azure-sdk-v1-3/

To get IIS logging and transfer to storage working

http://robindotnet.wordpress.com/2011/02/16/azure-toolssdk-1-3-and-iis-logging/

Known issues

http://msdn.microsoft.com/en-us/library/gg508668.aspx

The post Moving to Windows Azure SDK 1.3 appeared first on .

• Name has changed (windows live mesh 2011 <– live mesh <– windows live core (wlc) )
• Live mesh is not supported on windows phone yet
• Shell Integration seems to have been been removed (remember that “make folder live” !)
• News feed also seems to have been dropped.  News feature really lit up the live mesh client which felt boring with just “sync” till then (checkout the image of news feed here).  News feed was great and we quickly integrated news into live mesh mobile
• Dont see any apps now
• Integrated with skydrive

It looks like windows live mesh 2011 is solving the core sync problem.   Sync is one of the hard problems, especially in  a mesh scenario.  FeedSync was a very elegant sync mechanism for mesh sync.  From mobile perspective, it gets a bit more tougher given the storage, network connectivity, bandwidth and app-responsiveness issues.   Added to this, windows mobile had few limitations on shell support which made it difficult to track edits and deletes (lack of fileid).   Overall, feedsync was a very simple and elegant sync protocol and I am assuming it is used in mesh 2011 as well.

Here are some of the screenshots of the ‘old live mesh”

Old “device ring” of live mesh which still seems to be accessible (just took this screenshot from live site).

Here is s a screenshot of live mesh mobile.  I dont have this now, so took this from here)

-gopi

The post Windows Live Mesh 2011 appeared first on .

Enterprise mobility market was predominantly blackberry phones, to some extent windows mobile phones and ofcourse the laptops.  The IT department would choose the handsets that they would support and force it down on the employees.  But now, the employees are seeing a huge choice in front of them and want to play and use the new gizmos to improve productivity.

Enterprises are facing tremendous pressure from their employees to support new devices that they love than what the enterprise IT department “wants to support”.  Clearly, it is now going to be driven by the choice of the employees and enterprise IT departments have to figure out how to meet this demand of “choice” from their employees while keeping the system manageable and secure.

It is going to be interesting to see how the enterprises evolve their IT to meet these needs.  Some of the trends seem to be

• See this as a new experimentation phase, allow more devices.  This is a wise move as it helps in staying on top of the mobility changes and take advantage of choice
• Reduce expectation of security a bit as the upside on productivity is high. This is especially true of SMBs where security is not the most important factor.
• Move to “get your own device and we will support it in the IT”.  This helps in reducing the capital costs as well as the pain of managing the fast-changing mobility devices
• Move to SaS based solutions (from on-premise) which have a mobile component.  This helps in making most of the mobile devices without extra effort from the IT.

I hear from lot of my friends in the mobility space that enterprises are adopting tablets and smartphones at a very rapid pace.  Some of them are redesigning the tablets for custom applications which makes it easier for their mobile workforce.  Irrespective of which trends the enterprises latch on to, one thing is clear.  The enterprise mobility as we know has changed: changed for good, changed for a better choice and sooner the enterprises embrace it, the better.

The post Enterprise mobility appeared first on .

I have been involved in product development for the last 15 years and built various products from scratch.  One of the advantages as well as challenges of building a product from scratch or adding a sufficiently large feature is you get to choose the platform or platform components to build the product/feature upon.   Some of the real world examples that I was involved in are:

• Choosing the structured storage components, language (c++ or c#, STL or not, exceptions or not ),  and other dependencies  for live mesh mobile client
• Choosing the right platform for analytics work to munch upto peta bytes of data
• Choosing a networking stack and integrating it into ethernet products
• Choosing the right network processor or chipset for ethernet telecom products at tejas networks which needed to cater to evolving standards and features which were not clear at the time of design
• Choosing the platform to build www.citizensocial.com at Wignite

Though technical capabilities of the platform form a crucial input, more often than not, many other factors contribute to the final decision.   The final decision is always an engineering-style decision, which according to me is “choosing when there is ambiguity and trade-offs on many fronts“.  Assuming the possible options have cleared the basic requirements, it is often a gut call based on factors which may not necessarily be technical in nature.  Some of the factors are

• Technical comparison : which platform is better under various scenarios (speed, memory consumption, concurrency support, is it platform agnostic such as works on various flavors of *nix, say heat dissipation for chipsets (yes network processors used to be power guzzlers) etc )
• Platform Cost (cost of using the platform such as license costs, are there forces around commodotization of the platform which will drive its cost lower in future).
• Time to Market: This is extremely important.  We once had to choose between a chipset  or an in-house solution-to-be-developed for one of our crucial product of an OEM relationship.  The in-house solution  was perceived to be cheaper in the long run than the chipset in addition to building IP of the company.  But the tradeoff was that  the in-house development would take potentially 6 months more time and this would kill us in the market.  We went for the chipset based solution for that specific product and it turned out to be the correct decision (the decision was taken by one of my managers).
• Developer cost (ex: per-hour cost of a python developer is higher than say, a php developer)
• Support availability and support cost(how is the support from developers, how fast is it, if open source what is the number of active users on forums, cost of support)
• Technology trends (is the platform going to be around,  will it evolve to suit your requirements in future though it caters to only 80% of your requirements today,  do you see it getting obsolete).  Remember there used be things like google gears, google wave, yahoo pipes, intel and ibm network processors and many more .   This is very tough to guess though it looks very easy in hindsight.
• Product roadmap (can your platform choice support your product roadmap and evolution)
• Synergy with expertise within the company (If your developers are well versed with java, stick to it.  Dont move to c# just because there is one extra feature in C#)
• Other reasons such as close-tie ups (remember wintel combination)

My take on the selection process is

• Be very clear on the minimum requirements that is required of the product (ex: should scale to atleast 30million connections, unit price should not be greater than 5$ etc)
• Be un-emotional about technology choice and beware of especially the NIH (not invented here) syndrome
• Do a technical due diligence (it need not take months together.  Often, you can get a good sense in a week by writing some test code/ reading the datasheets)
• Discuss and solicit critical feedback from your peers and managers.. assuming your managers are good
• And last but probably the most important : apply common sense.  Often one can see where a platform is going by just thinking clearly about it and applying yourself to the problem.  Don’t pick up a platform just because some big gun is shouting at the top of his voice that their platform is the best choice.

It is often not a easy decision to choose a platform for your product but a very exciting and challenging phase of product development.  One learns a lot during the process of selection as well as later when we get to see the affects of our decisions and learn from the mistakes.  I have been lucky to have had excellent managers and rock star team mates who helped in driving the decisions by their critical question and valuable insights.

Do you guys have any thoughts or other interesting experiences to share?

–gopi

The post Choosing a platform for your product appeared first on .

Just realized that it is a silverlight problem on chrome.  moved to IE and cut-paste works again.

This is a tip just in case somebody is having trouble with this issue and didnt realize that it is a browser+silverlight issue

The post silverlight cut-paste doesnt work in chrome appeared first on .

“Build 6002: This copy of windows is not genuine”

obv it is an internal build with changes for data center but it is funny to see the message that says “microsoft is not using genuine windows in their own data centers”

ms-datacenter-piracy

The post Windows OS pirated in microsoft data centers :-) appeared first on .