Multi-Factor Authentication Now Included in Standard Office 365 Subscriptions

Office 365 Multi-authentication is now available for all their standard plans at no charge.

The “multi” aspect of the authentication is in addition to your username and password, you can require that your users also acknowledge a phone call, text message, or an app notification on their smart phones before they can sign-in successfully.

mfa_01

Multi-factor authentication can be added for each individual user in your organization as required. Once the user is assigned, they will be asked to configure their second authentication mechanism.

Read More

How Much Data Can PowerPivot Really Manage? How about 122 Million Records?!

Excel 2013 is an amazing tool in terms of its increased scalability and its ability to handle large data sets.  Combined with PowerPivot, PowerView and Power Query, Excel 2013 is now Microsoft’s leading tool for business intelligence.

What is amazing is how much data Excel 2013 can process in memory – no need for cubes, data warehouses, centralized servers, etc.  if you can just load some large data files straight into Excel. 

When we hear about “Big Data”, how big is big?  Just using Excel 2013 and some raw CSV files, I can load quite a lot of data. 

14 Million Records

Earlier I wrote a blog post where I loaded approximately fourteen million records from the City of Toronto representing every parking ticket issued in the past several years.  Using this dataset loaded into PowerPivot, I can use PowerView to chart the volume of parking tickets by day of the week, by year or by month.

image3

Loading up 14 million records in Excel 2013 takes about 3 minutes to completely refresh the data source from the original data source in Power Pivot.  Once the data is loaded, the Power View refresh is fast enough to be highly usable.

122 Million Records!

In looking for large data sets, I found the publically available data of every airline delay in the United States from 1998 to 2008.  This data is available from the Bureau of Transportation Statistics.  I downloaded all of the CSV files and dropped them into a single directory.  There are approximately 122 million records and over 11 gigabytes of data across the entire set of files.

This dataset has been thought of as “Big Data” historically and has been used for testing both generic statistical analysis techniques as well as specifically for trying to optimize the prediction of delays based on factors such as weather, origin of the flight, etc.

As an experiment, I tried loading the entire dataset into Excel 2013.  To load multiple files into Power Query, you can choose a folder location and pick up all the files in that folder.

If you then click on the expand arrows to the right of the Content column, you can select the actual contents of the files to be imported instead of just the file metadata.

You can then Apply and Close and let the import begin!

image5

Here are some interesting performance stats while my machine was chugging away:

image

image

image

As illustrated by the graphs, performance is dependent on two key things: 1) available RAM and 2) available disk IO.  CPU is barely being taxed.  This intuitively makes sense – essentially this is a very large loading job from disk into RAM.

The results are that it worked – I was able to import 122 million records into Excel.  Here is my PowerPivot table with 122 million records!  From beginning to end, it took approximately 1 hour to load the entire data set into memory.

Is the Data Usable?

I was able to load the data, but is usable?  Here are a few quick tests to see if I could actually manipulate the data in some meaningful way.  While not an exhaustive test, here are a few basic usability tests I tried to determine whether Excel could do some basic reporting on the dataset.

Loading and Saving the Excel File

After loading up all this data, could I actually save the Excel file and could I reload the file from disk?

Saving the file to disk was quite fast – it took about 2 minutes to save the file.  The file itself is about 2.2 gigabytes – there must be some compression going on here because the original files were more than 11 gigabytes.

Loading the file was very fast – it took seconds.  However, when I tried later after my computer went to sleep and I restarted Excel, it was significantly slower – about 2-3 minutes.

Linking Data to Other Tables

As part of the original dataset, there was also supplied three CSV files that had data for Airports, Carriers, and Planes.  I loaded each of these into my model.  In addition, I added a Date table from Azure Marketplace.  I linked these into my PowerPivot model as relationships.

Most of these were quite quick to create – within a minute.

Create a Calculated Column

The Airline Delays dataset provides columns for Year, Month, DayOfMonth, and DayOfWeek.  So I created a column to convert the Month as a number to the Month name (e.g. 2 = February).  I also created a column to extract the hour from the Departure Time column.

Running these basic calculations was pretty fast – about a minute.

Creating a Power View

I tried to create a Power View sheet with the Delay Hour field and the DepDelay field to show how departure minutes are related to the time of the day.  This was VERY slow – it took minutes to refresh the data every time I made a change.

This is where having a pre-aggregated cube in SQL Server would make things much faster – having everything aggregating dynamically is slow when dealing with volumes this large.

But it works!

Conclusion: Pushing the Limits of an In-Memory Based Database

14 million records was useable – I can manipulate the data, save the spreadsheet, load it into a Power View and it works quite well.  It’s not fast but it’s usable.  I could imagine doing analysis against this data pretty easily.  I can do aggregation really easily against dates to filter the data by year, month, day, etc.

At 122 million records, Power Pivot shows its limits.  As everything is loaded into memory dynamically with no pre-indexing, once my laptop uses up its available RAM, it becomes slow to process the data.  Working with PowerView also becomes challenging because every time you change the configuration of a chart or view, it takes minutes to recalculate the result.

Read More

Yet Another Example of Microsoft’s New Open Source Commitment – OpenDaylight

OpenDaylight is an open source framework to enable Software Defined Networks (SDN).  SDN is seen by the industry as a more agile approach to delivering networks that adapt to the requirements deployed within that network.  For a dynamic infrastructure such as a massive cloud, having a highly adaptable network becomes a key piece to managing the infrastructure.

The OpenDaylight Project has garnered the attention of the industry and is being backed by the Linux Foundation.  Key networking players such as Brocade, Cisco, and Juniper are working together on this common platform. 

Microsoft is a platinum member and has ten full time developers contributing code to the project. 

In a similar way to their participation in the Hadoop project, Microsoft seems to have figured out that driving the industry forward in some of these key technologies is going to take multiple organizations, lots of sharing across the industry and pooled knowledge.  Microsoft also seems to have defined openness as a competitive differentiator compared to Apple, Google and Amazon.

This is a massive departure from the Microsoft anti-open source era of the early 2000’s where Steve Ballmer described Linux as a “cancer that attaches itself in an intellectual property sense to everything it touches”.

Read More

Constructing a Basic Cost Comparison Between On Premise and Office 365

We have lots of customers who want to understand the difference in pricing between what is offered as part of Office 365 and what they could have on premise.  The challenging answer is that Microsoft Licensing requires a massive degree of expertise, discount negotiations on both sides and alignment of different features available in different packages.

Most Microsoft on premise products have a combination of a server license and a user CAL license.  For Office 365, the licenses are monthly subscription priced on a per user basis.

So the key questions that need to be answered to create a comparison model are:

1. What features and services are we including to compare (e.g. Office, SharePoint, Exchange, etc.)?

2. How many users are we trying to support with these services?

3. Are there any extraordinary capacity needs, e.g. additional performance, storage, encryption, security, integration, etc.?

One of the Microsoft Cloud sales staff has put together a cost comparison excel sheet that can be used as a tool to compare different license configurations and scenarios.

As an example scenario, I used the spreadsheet to calculate a typical E3 Enterprise Office 365 scenario.  The Office 365 E3 subscription includes:

  • Subscriptions to Microsoft Office for up to 5 desktops
  • Email via Exchange
  • LYNC
  • SkyDrive Pro
  • Public Website Hosting
  • Spam and Malware Protection
  • SharePoint with Enterprise features such as BI, eDiscovery, and Social

Pricing Out the E3 Subscription for 300 Employees

For an organization of 300 employees @ $23.20 / month, the cost for all of that is $83,520 per year.

Let’s now price out the equivalent licenses for SharePoint, LYNC, SQL, Exchange and Office for 300 employees.  Let’s assume you have high availability setup and SQL running in high availability (since this is what you are getting as your cloud service).

CAL Licenses

For my 300 employees, I need the following CAL’s:

  • Exchange Standard + Enterprise
  • SharePoint Standard
  • Office Pro Plus
  • LYNC Standard + Enterprise

All of these should be priced with Software Assurance included since we get automatic upgrades as part of Office 365.

For 300 users, this is equivalent to $418,500 in one time costs at list prices. 

Pricing the Servers

A minimal high availability SharePoint farm is 6 servers: 2 web front ends, 2 app servers, and a SQL cluster with two nodes.  Each one requires Windows.  SQL is also required for the application servers to run Excel Services, SQL Server Reporting Services and PowerPivot Services.  Assuming each has a couple cores, our SQL costs get pretty expensive pretty quickly. 

LYNC requires similar front end servers and a back end SQL database for storage.  With high availability, we need another SQL cluster in the back end. 

Exchange requires client access servers and mail box servers. 

The Results of my Sample Comparison

The result of my high level license configuration was a chart that looks like this:

image

Note that this is just server and CAL licenses, and does not include any hardware pricing, storage pricing, services to maintain the infrastructure, backup and restore infrastructure, monitoring, etc. 

Conclusion: It All Depends on Your Needs

In doing a proper analysis of Office 365 and On Premise deployments, the cost comparison requires some deep comparison of a complete picture on both sides including software licenses, hardware, infrastructure, data center costs, monitoring, services, etc.  Depending on your organizational needs, the specific cost picture will vary considerably.  In addition, your organization will have different license terms with Microsoft depending on your size, your industry and geography.

Have a look at this pricing tool as a starting point…but in general look at the total cost of ownership of your Microsoft investments and you may find that an Office 365 subscription model represents a significant savings over an on premise implementation.

Read More

2014 Olympic Winter Games are Streaming Through Windows Azure Media Services

NBC uses Microsoft Azure as its streaming platform for the Winter Olympics.  They use Windows Azure to deliver live streaming through its Windows Azure Media Services.

NBC on Windows Azure

Windows Azure Media Services is an on-demand, cloud based service that is for encoding, encrypting, transforming and serving video assets.  Like all cloud services, Windows Azure Media Services provides the power of on demand pricing so this is a perfect model for event based programming where there is a massive spike in demand for video processing services. 

Read More

Are you a Startup? You should join BizSpark!

Are you a start up with less than $1 million in revenue?  Microsoft is interested in supporting your success and hopefully getting you to adopt their software as you grow into a successful company!

Microsoft has a specific program called BizSpark for startups with less than $1 million in revenue.  In joining the FREE program, you get access to the equivalent of MSDN Ultimate which provides your development team with all the software they need for development/testing purposes (including Visual Studio, Team Foundation Server, SharePoint, SQL, etc.) as well as $100 a month in Azure credits. 

Read More

For Those Who Need MASSIVE Computing Power, Microsoft Offers New Higher Performance PaaS Instances

Microsoft has just announced that as part of its Platform as a Service, you can now run compute intensive applications on some new virtual machines that come with 8 or 16 cores with 7 gigabytes per core!

The price for one of these powerhouse machines is $1923/month for the 8 core instance or $3845/month for the 16 gig instance.

In addition, Microsoft is providing a massive 32 Gbps InfiniBand network for low-latency and high throughput communication.

Note that these instances are only currently available in US North Central and Europe West locations. 

While available only through Microsoft PaaS instances, these new servers will be available as IaaS virtual machines in the future.

These instances are targeted to high performance computing applications such as weather forecasting, financial analysis and energy exploration. 

Microsoft also provides an entire high performance computing including an implementation of the Message Passing Interface standard used for high performance parallel computing.

Microsoft’s commitment to both public cloud and hybrid cloud architectures means there are different ways to create high performance computing clusters:

Hybrid cluster

Hybrid High Performance Computing Cluster

Cloud-based cluster 

Cloud Based Higher Performance Computing Cluster

Read More

Which Version of SharePoint is Right for You?

With SharePoint 2013 and Office 365, there are now 15-20 different licensing options when it comes to implementing SharePoint.  The master service description is available on TechNet and it provides in detail the differences in licensing options in terms of the features of SharePoint that are available. 

image

A couple of consultants have taken this information and created filterable excel files that can be used to compare versions in detail.

Thank you gentlemen for putting these together!

Read More