Tag Archive for Cloud Computing

Microsoft’s Cloud-Scale Big Data Solution

The rate of new Azure services designed to address Big Data type problems continues to accelerate.  This is due in large part to the continued maturity of Azure as a stable and reliable Public Cloud offering.  Indeed, the key to profitability in a business data science effort has much more to do with the ways in which today’s cloud services deliver big data capabilities cost effectively.

Several pieces of the Microsoft Big Data solution are delivered in Azure, allowing us to truly build Big Data solutions at “Cloud-Scale”.  In the context of the 3-V’s of Big Data (Volume, Variety and Velocity), “Cloud-Scale” means massive cost-effective storage, schema-less data and “ingestion” of millions of rows of data per second.

1. Azure Blob Storage eliminates the need for your Data Science team to provision a Petabyte or more of redundant storage for your Data Lake.

2. Azure Service Bus and Event Hubs deliver telemetry ingestion from websites, apps, and devices.  Intake millions of events per second from a wide variety of sources.

3. Azure Stream Analytics handles the transformation of data “In Motion”.  In traditional BI, we aggregate and summarize data “At Rest”.  The velocity of Big Data requires technology which aggregates and summarizes data in motion.

4. Azure HD Insight is Microsoft’s Apache Hadoop distribution.  Developed by Hortonworks, its 100% compatible with Hadoop toolsets such as Pig, Hive, SQOOL, etc.  The Map and Reduce components can be deployed as Microsoft.NET assemblies, written in Microsoft C#

I’ll add another V, because “Cloud-Scale” does not translate well in the land of humans.  In order for insights to be actionable (and profitable), our Big Data solution must simplify the information Visually for mere mortals.

5. Power BI is a cloud service that lets you share, collaborate and access your Excel reports anywhere on any device.

The Big Data picture is coming into focus and it does not require a legion of consultants, hardware, or Hadoop experts to achieve.  Enterprises which have invested in Microsoft Business Intelligence can move into the Internet of Things era with a mid-size data science project, and grow to “Cloud-Scale”.

Deploying Software with Windows Intune

New trends in IT are emerging with the biggest being the move for businesses to cloud services. One of the newer cloud based services from Microsoft is Windows Intune. Intune is Microsoft’s cloud answer for on-premise management based technologies such as System Center. With Intune you can manage windows updates, mobile/tablet devices, perform hardware/software inventories and manage antivirus. In addition, you can now choose 2 plans for Intune with or without a Windows 8 Enterprise license so you can be sure your desktops/laptops are running the latest version of Windows.

In this blog article, I’ll be walking you through deploying a 3rd party application, Adobe Reader by using Windows Intune. This is a great feature from Windows Intune giving IT the ability to deploy 3rd party applications with a few clicks versus having users themselves install the software or IT walking around to each desk. Time is money these days in businesses and businesses are looking at where they can save money and increase productivity. In my next article, I’ll discuss on how to deploy an Adobe Reader update so you can be sure everyone running 3rd party applications are kept up to date.

To get started, if you don’t already have the Adobe Reader MSI file, you can download it here on Adobe’s FTP site: ftp://ftp.adobe.com/pub/adobe/reader/win/11.x/

Once the file is downloaded, log into your Intune Admin console and click on the Software icon on the left hand side of the screen.

1. Click Add Software under Tasks on the right.

clip_image002

2. Run the application and sign in with your Intune credentials. Click Next.

clip_image004

3. Keep default selections and browse to your downloaded MSI file from the FTP site. Click Next.

clip_image006

4. Enter a publisher name. You can also upload an icon for the package. (Optional). Click Next.

clip_image008

5. Keep the defaults and click Next.

clip_image010

6. No command line arguments needed for Adobe Reader. However, note you may need to configure silent installs for other applications. Click Next.

clip_image012

7. Click Uploadon the summary screen and Adobe Reader’s .msi file will be uploaded to your Windows Intune storage.

clip_image014

8. Once the upload is complete. Click view software properties to go directly to the deployment settings for the software package.


clip_image016

9. Select the devices/users you’d like to deploy the software to then click Next.

10. Specify the type of deployment whether is it required or available to install then a deadline. Click Finish.

If you’ve followed all of the above then you should have successfully deployed Adobe Reader through Windows Intune. Comment below if you have issues or questions.

Data Warehouse as a Service is Here

Cloud Computing has a new acronym: WaaS, or Warehousing as a Service, and if you are intrigued by the cost savings promised by Infrastructure as a Service, Warehousing as a Service will take your breath away.

BitYota has been a prominent start-up in the race to provide ultra low-cost data warehouse capabilities in a public cloud infrastructure.  In order to pull off a BI solution in the cloud, a vendor has to be able to provide a very compelling price point, both on storage and on compute processing.  At $1500/month, the BitYota “Starter” plan looks like a good fit for small to mid-size companies.

image

 

Amazon Web Services has upped the ante with their announcement last week of RedShift Data Warehouse service.  Designed to be easy to launch with a few clicks from the AWS Management Console, Redshift is priced at under $1,000 per terabyte per year, but customers will have the ability to scale up to a petabyte or more of storage space.

That’s right: $1000 per TB/year.

RedShift is currently in beta; expect the full release in early January, 2013.

Microsoft’s plans for supporting WaaS on Azure are unclear.  The concept has been incubating within the SQL Azure team since the beginning.  Windows Azure does supporting SQL Database Federations, groups of SQL Databases which implement Horizontal partitioning, or Sharding.  But it is unclear how to go about pricing this: unstructured data is extremely cheap, but when you need structured data, the price does not yet compete with AWS or BitYota.

Is the Enterprise Data Warehouse dead?  Will it be in 5 years?

Cloud Databases Gaining Momentum

Typical cloud computing analysis over the past few years has focused upon 3 fundamental offerings: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS). Now, a new class of cloud offerings, dubbed Database-as-a-Service (DaaS ??), has emerged in response to the high costs of provisioning, administration, security, backup and recovery of on-premise data management solutions.

Cloud Databases are not simply IaaS with a database license. In order to be competitive in the market, cloud database vendors must provide an abstraction layer and standardization which essentially removes the requirement to have a database administration role on your staff.

Fair enough, but if you already have an on-premise database architecture in place, why would you want to put your nextdatabase in the cloud?

According to Forrester Research, there are five prominent use cases:

1. Application Development and Testing – application developers and testers rarely get the luxury of comprehensive stress testing, because it is too expensive to build out anything other than a basic database to test functionality. Cloud databases offer quick ramp up time for testing of high load applications – and only for a small duration of time, attractive in a pay-as-you-go price model.

2. Mobile Applications – mobile applications which compose data from a variety of applications and services need a central repository, or hub, from which to pull the data they need. Cloud Databases offer a perfect single data source to aggregate data from multiple sources.

3. Departmental Collaboration – collaboration software has become an absolute necessity, yet many departmental applications find it difficult to justify enterprise collaboration tools and spend. Cloud databases are a lightweight spend for lightweight data accessed globally.

4. SMB and departmental applications – smaller organizations have many of the same enterprise data needs as larger enterprises, with a fraction of the IT budget. I can almost hear the Access database developers groaning.

5. Database backup and archive – enterprises have a variety of data retention requirements which can be met with a cloud database more efficiently than on-premise. How much are you paying to hold on to historical data which you rarely access?

I think I can distill these five points into a business no-brainer. Cloud Databases are all about speed. Not speed of access to your data, but speed from business concept to marketplace.

Another important note, Forrester’s analysis identifies four leaders among cloud database vendors, with a pack of contenders nipping at their heels. This indicates to me that the market is maturing rapidly, with well understood selection criteria evolving for CIO’s looking to make a selection.

The leaders include:

  • Microsoft’s Windows Azure SQL Database, a traditional SQL Server data store in the Microsoft cloud.
  • Amazon Web Services (AWS) RDS – which is provides hardware infrastructure supporting several RDBMS platforms: MySQL, Oracle and SQL Server
  • AWS DynamoDB – a NoSQL implementation designed for holding unstructured Big Data
  • Salesforce.com’s Database.com – this is running the data platform for SalesForce itself

Check out the full analysis here:

http://www.forrester.com/pimages/rws/reprints/document/70541/oid/1-LLP641