Clarifying “Windows PE”

Uncategorized No Comments »

Hi All,

Today I wanted to clarify a commonly used term “Windows PE”,which some of the people always have doubts on.

Lets see how many of us can get it right !!!

Windows Preinstallation Environment (Windows PE) 2.0 is a minimal Win32 operating system with limited services, built on the Windows Vista kernel. It is used to prepare a computer for Windows installation, to copy disk images from a network file server, and to initiate Windows Setup.

Windows PE is not designed to be the primary operating system on a computer, but is instead used as a standalone preinstallation environment and as an integral component of other setup and recovery technologies, such as Setup for Windows Vista, Windows Deployment Services (Windows DS), the Systems Management Server (SMS) Operating System (OS) Deployment Feature Pack, and the Windows Recovery Environment (Windows RE).

In the past, MS-DOS-based boot floppy disk was used to start a computer and then connect to a network share where a Windows installation source or disk image was located or to troubleshoot and recover a copy of Windows that did not start.
But we all know how many limitations it had !!!!!
-No support for the NTFS file system.
-No native networking support.
-No support for 32-bit (or 64-bit) Windows device drivers, making it necessary to locate 16-bit drivers.

So now we have Windows PE to solve all those limitations and offer great features like
-Native support for NTFS 5.x file systems, including dynamic volume creation and management.
-Native support for TCP/IP networking and file sharing (client only).
-Native support for 32-bit (or 64-bit) Windows device drivers.
-Native support for a subset of the Win32 Application Programming Interface (API); optional support for Windows Management Instrumentation (WMI) and Windows Script Host (Windows SH).
-Can be started from different kinds of media, including CDs, DVDs, USB flash devices (UFD), and Windows Deployment Services (Windows DS).

Why Make the Move to Windows Server 2008

Uncategorized No Comments »

Hello all,

Finally, we are at a stage when companies are ready to upgrade to Microsoft Windows Server
2008.   Yet, some people still want specific reasons as to why exactly Server 2008 is better than Server 2000.

Here are some of the highlights that are causing us to jump.

With Hyper-V, the Windows Server 2008 virtualization solution, a single physical server can host the workloads of multiple Line of Business servers. Hyper-V helps organizations to achieve optimal use of their hardware resources and provides the agility needed to adapt to changing IT needs.

The Server Manager Console provides a single, unified console for managing a server’s configuration and system information, displaying server status, identifying problems with server role configuration, and managing all roles installed on the server.  It allows administrators to complete tasks with fewer clicks without having to navigate between multiple tools and interfaces.

3. Subsystem for UNIX-based Applications (SUA)
Windows Server 2008 includes Subsystem for UNIX-based Applications (SUA), a multi-user UNIX environment that supports more than 300 UNIX commands, utilities, and shell scripts. Users  can maintain one user name and password for Windows domains and UNIX systems, synchronizing the credentials automatically when one changes.

4. Active Directory Rights Management Services (AD RMS)
Companies need to share information with partners and clients without losing control over that information. Rights Management Services enables organizations to control how documents are used—including who can view them, whether they can be printed, even whether they can be forwarded or deleted—both internally and externally.

5. Server Core
The new Server Core installation option provides a minimal environment for running specific server roles. This helps improve reliability and efficiency, giving the IT department the ability to better utilize existing hardware. It also simplifies ongoing administration and patch management requirements by reducing the need to update unneeded files and functionality.


IPv6 – Part 3

Uncategorized No Comments »

Global Unicast, Link-local, Site-local, Unique-local, Multicast
In addition to the Global Unicast addresses that start with either a 2 or 3 and the reserved “::1” for localhost, you will need to recognize and respond to: FF, FE80, FEC, FC, and FD.

Multicast – FF
First, there is no Broadcast in IPv6 and that will take a while to adjust to – the lack of broadcast. In IPv4, when a host system did a “DHCP Discover”, the broadcast had to be processed by all the systems, not just the DHCP servers. To reduce the performance bleed of systems, IPv6 uses the more targeted Multicast. Any IPv6 address which starts with FF is a multicast address. Example – FF02::1:2 is reserved for DHCP agents (server or relay agent) – RFC 2375.

NOTE: With IPv6, there may not be a need for DHCP, as hosts have the option to auto configure themselves using the new ICMPv6 router solicitation/advertisement (router discovery).

Link-local addresses – FE80::/10 (FE8, FE9, FEA, FEB)
As mentioned earlier, the most common support call is: What’s an FE80? Well, with a small number of systems, it’s always been convenient to be able to just plug everything together, power-up, and have all the systems find each other and start communicating.

With IPv4, the first part of the configuration (host id) can be provided by APIPA – Automatic IP Addressing.  When a system doesn’t receive a response to its request for automatic address assignment (DHCP Discover), it’s allowed to generate an address – in the form of: 169.254.random. Advantage – all systems on the same segment (link) can generate a unique ‘random’ address and start communicating. Limitation:  packets are not routed – communication is limited to the local segment (link).

With IPv6, this is referred to as a Link-local address and takes the form of – FE80:/10. So, in a small, simple environment (single link/segment), there’s instant communication. However, in a normal corporate environment (multiple, routed, networks), like APIPA, you weren’t assigned a “useable”, routable address. And you can answer the question, “What’s an FE80, and why can’t I get to the server?”

Site-local – FEC0::/10 (FEC, FED, FEE, FEF)
When it became obvious that IPv4 would soon run out of addresses, NAT (network address translation) was there to save the day. Some of the still available addresses (not the most attractive) were set aside and labeled as “Private”, with normal (routed on the Internet) addresses labeled as “Public”. The Private addresses would not be routed by Internet routers (never appear on the Internet), so they could be used over and over by any company. NAT would provide the Public (external) to Private (internal) translation.

In IPv4, the Reserved Private addresses are: 10/8, 172.16-31/12, and 192.168/16 – routed within a company, but intentionally not routed between companies.

In IPv6, these are referred to as Site-local and use addresses in the range of FEC to FEF or FEC0::/10. Example: FEC0:0:0:FFFF::1 is the default assumed address for a system’s internal DNS server.

I would like to say that we’ve finished your crash course on IPv6 addressing and special addresses, but there’s one more – the newer Unique-local, that officially deprecates Site-locals.

Unique-local – FC00::/7 (FC and FD)
RFC 3879 …”As currently defined, site local addresses are ambiguous: an address such as FEC0::1 can be present in multiple sites, and the address itself does not contain any indication of the site to which it belongs.”…

Unique-local addresses serve the same purpose as Site-Local, but made provisions for a “Unique” Private address.

The upper 64 bits (network address) are broken down into 3 fields: the first 8 bits would identify the address as a Unique-local, then the next 40 bits would be a ‘global id’, with the final 16 bits being a ‘subnet id’. The ‘global id’ could identify a specific company, then the ‘subnet id’ could be used to identify “sites” (locations) within a company.

The top 7 bits identify the address as a Unique-local. Bit 8 is the difference between FC the ‘managed’ addresses and FD ‘unmanaged’ addresses. With FC, the 40-bit global id could be assigned (managed) by a future organization (e.g. ULA-Central). With FD, companies simply use a pseudo-random algorithm to generate the 40-bit global id. With either system, a company would then use the 16-bit subnet id to identify their internal sites (e.g. geographic locations). Every internal network within a large world-wide corporation can be uniquely identified for reliable routing.

Pop quiz:
Given a multiple choice question of:
A:     FF01
B:    2001
C:    FE80
D:    FD92
E:    FEC0

Which is a Multicast address?

Which is a Global Unicast address?

Which is a Link-local address?

Which is a Unique-local address?

Which is the deprecated Site-local?

Will ’7′ Be Windows’ Lucky Number?

Uncategorized No Comments »

We’re taking a short break from IPv6 series to talk about the new Windows 7.0, Microsoft’s new operating system that many are looking forward too.

With download problems solved, no killer problems seem to have arisen so far in Win7′s public beta.  Suprisingly.

In fact, overall, the process appears to be going overall fairly smooth up until now, with the exception of what happened on Friday.

Microsoft (NASDAQ: MSFT) servers were so overwhelmed even before the Windows 7 public beta was slated to be posted on Friday. A crush of traffic even before the beta build was released forcing Microsoft to deploy more servers than originally planned, which ultimately delayed the program’s start until Saturday.

After using it for a day or two, people didn’t have any complaints, which was refreshingly good news.  Perhaps a dozen users, including one who claimed to be a Microsoft employee, did report that, for them, the beta freezes when they try to run Windows Live Messenger 2009.

We have high hopes from Windows 7.0.  Let’s keep our fingers crossed.

IPv6 Blog Course – Part 2

Uncategorized No Comments »

Time for some technical details:

Colon Hexadecimal notation, Zero Suppression, and Global Unicast addresses
What to do with these 300 trillion, trillion, trillion addresses?

We should reserve the biggest chunk to address the biggest issue – Global/Public (routed on the Internet), “unique in the world” addresses. One eight (1/8) of the addresses have been reserved for IPv6 Global Unicast addresses. If the top 3 bits of the address are “001”, then it’s a Global Unicast address. The top 3 bits (any 3 bits) can store 8 combinations (23), with 001 being one of the 8. But to a human, not a computer, this would look like either the number 2 or 3. Why?

Colon Hexadecimal notation
IPv6 uses a colon (:) hexadecimal notation, not the doted (.) decimal notation used by IPv4

IPv4 example —
IPv6 example — 2001:0db8:0000:0000:0000:002e:0370:2334

IPv4 cut the 32 bits into 4 octets (8 bits), each represented as a decimal number. The choice of decimal seemed like a good idea at the time (the logical choice), but experience showed this not to be true. When troubleshooting, it helps a lot to be able to see things as the computer sees them – in binary. But you don’t always have a decimal to binary conversion calculator handy. We found that in training, we were spending hours, if not an entire day, showing different techniques that could be used for reliable conversion from decimal to binary and back.

In the early days of computing, it was common to use non decimal based number systems, such as octal, hexadecimal (hex), and of course binary – number systems based on a power of 2. With these number systems, it’s a simple skill to convert a number to binary and back. MAC addresses have always used hex. IPv6 is returning to a more logical number system – hex. To most people, a number like FE80, looks a little strange at first, but in short order, most adapt start taking advantage of the ease of conversion – hex to binary and back. Hex gives us a human friendly representation of binary coded values.

Hex uses four bits for each number – 24=16 (hexa decimal). These four bits are also referred to as a nibble – two nibbles to a byte. The characters used are 0 to 9 then A to F for the additional six. Zero would then be 0000, One is 0001, Two is 0010, Three is 0011, with F being the last combination – 1111.

IPv6 breaks the 128 bits into 8 blocks – each block then represents 16 bits, expressed as 4 hex characters. It’s definitely a longer number than an IPv4 address, but with hex, it’s a consistent number of characters and more importantly, easy to convert all of part of the address to binary.

Zero Suppression
It’s common for an IPv6 address to have a lot of zeros in the middle or leading zeros in a block.
Example: 2001:0db8:0000:0000:0000:002e:0370:2334

Zero suppression allows us to first, simply eliminate any leading zeros in a bock.
Example: 2001:db8:0:0:0:2e:370:2334

Contiguous blocks of all-zero can then be collapsed and represented with a single double-colon (::)
Example: 2001:db8::2e:370:2334    ! Double-colon can be used only once – to avoid ambiguity.

The double-colon, by itself (::) is used to represent a special “no address” as opposed to an address of 0.

“::1” is reserved for troubleshooting – being the ‘loopback’ address, equivalent to IPv4’s This can be seen clearly by typing: Ping -4 localhost, then Ping -6 localhost.

Global Unicast Address
2001:0db8:0000:0000:0000:002e:0370:2334 or 2001:db8::2e:370:2334 is an example of a Global Unique address. The computer determines this by looking at the top three bits of the address and the value of 001. The hex number of 2 is seen as 0010; therefore a Global address. But the hex number 3 is also seen as a Global address – it has a value of 0011, with the top three bits also a 001.

Any IPv6 address that starts with a 2 or 3 is a Global Unicast address; will be forwarded by Internet routers and is a “unique in the world” – the equivalent of IPv4 Public addresses (non Private).
NOTE – 2001:db8 has been reserved for documentation purposes – RFC 3949.

Global Unicast, Link-local, Site-local, Unique-local, Multicast’

In addition to the Global Unicast addresses that start with either a 2 or 3 and the reserved “::1” for localhost, you will need to recognize and respond to: FF, FE80, FEC, FED, FEE, FEF, FC, and FD.

Multicast – FF
First there is no Broadcast in IPv6 and that will take a while to adjust to – the lack of broadcast. In IPv4, when a host system did a “DHCP Discover”, the broadcast had to be processed by all the systems, not just the DHCP servers. To reduce the performance bleed of systems, IPv6 uses the more targeted Multicast. Any IPv6 address which starts with FF is a multicast address. Example – FF02::1:2 is reserved for DHCP agents (server or relay agent) – RFC 2375.

NOTE: With IPv6, there may not be a need for DHCP, as hosts have the option to auto configure themselves using the new ICMPv6 router solicition/advertisement (router discovery).

Link-local addresses – FE80::/10 (FE8, FE9, FEA, FEB)

Site-local – FEC0::/10 (FEC, FED, FEE, FEF)


IPv6 Blog Course – Part 1

Uncategorized No Comments »

IPv6 is normally described as simply evolutionary, but it’s both evolutionary and revolutionary. IPv6 is a fairly simple to understand evolutionary technical step for the network, but will probably turn out to be a revolutionary step for the industry.

The obvious reason that we need a replacement for IPv4 is the fact that we are running out of addresses – at a frightening pace. The U.S. continues to consume addresses at an increasing rate. But the really scary part is that the rest of the world (e.g. Russia, Ukraine, China) are just arriving at the party. They’re really thirsty and the punch bowl is empty – there are just a few sips (addresses) left in the bowl.

IPv4 is based on 32 bits, with 232 being about 4 billion addresses. This seems like a lot of addresses and because of this, in the early days a lot of addresses were wasted (e.g. Classful addressing, Class A, B…). If it weren’t for human creativity: classless addressing, the development of ISPs that charge increasing rates for the remaining Public addresses and the even more significant, successful use of NAT (Network Address Translation), we would be out of addresses and the party would be over. NAT allowed us to reserve the last of the addresses for ‘reserved Private internal’. These Private (not routed on the Internet) addresses could be used over and over by every company.

NOTE:  NAT did indeed save the day, but there’s really no long term future for NAT. It works fine for desktops wanting to reach a server on the Internet or even publishing a Web or Email server to the Internet. But it falls apart when you try to use it for server-to-server or desktop-to-desktop.

Operating Systems (O/Ss) and File Systems, based on 32 bits were also running out of addresses. The first to respond were file systems (e.g. NTFS) that moved to 64 bits with operating systems and their underlying hardware next to follow. With 64 bits, we have 264, which is 16 Exabytes.
But I and most people have a hard time trying to relate to an Exabyte. Is it bigger than a breadbox, bigger than a car, bigger than…? I do know that it goes from Mega, to Giga, then Tera, but there’s Peta before we get to Exa. I can relate to a Tera (because I just bought one), but it gets a little fuzzy at the Peta and above. I found that it was helpful to think of it as 17 billion, billion or 17 billion gigabytes.

IPv6 could have also moved to 64 bits, but with all the trouble it will be to deploy, they didn’t want to be guilty of another under-estimation – IPv4’s 32-bits. It was decided that IPv6 would move to 128 bits and skip the obvious 64 bits. After working with IPv6’s long addresses (very long), I had some second thoughts. I and others began to think that maybe 128 bits was ‘over the top’ and 64 bits would have been a better (less awkward), adequate choice. But then I got it.
We still, as before, need to split the address into two parts for network routing. With IPv4, it was a network id and host id, with the mask identifying the network id. IPv6 will continue to use a mask to identify the network id. The remaining bits for IPv6 are technically the ‘interface id’ as opposed to a ‘host id’ – start thinking about systems with multiple addresses, instead of just ‘the host address’.

We still need a mask, but we’re not going to need a lot of fancy masks or what I call a non-standard mask (e.g. /27 instead of /24). The creative borrowing of hosts’ bits (sub netting) was triggered by the decreasing availability of addresses and networks.
For example, a company might only be able to afford to rent a Class C subnet and not the entire Class C. So the ISP would supply a non-standard (subnetted) mask; plus the company sometimes needed to split (subnet) a network into multiple buildings. But NAT (reserved internals) was there to save the day – with a fresh supply of addresses; most companies were able to return to the preferred standard masks for the internal networks.

For IPv6, the obvious and most common mask will be a simple /64 – 64 bits for the network id and the remaining 64 bits for the interface id. This means that, even with a simple mask, we could assign network addresses to 17 billion, billion companies (or even houses and people) and then each company or individual would have available 17 billion, billion host addresses. The move to 64 bits, split into two 32 bit pieces, would all too soon, come up short again…

So the move from 32 to 128 bits (skipping 64 bits) was probably a wise choice. With 128 bits, how may addresses do we really have to play with – even before we do a little non-standard masking?

With IPv6, the 128 bits gives us 2128 ,  which is 340 Undecillion possible addresses (or 3.4 x 1038), like an Undecillion means anything to me … sounds a little silly to me, maybe an ‘un de silly on’ – this is even more challenging than the Exabyte (like an Extra big byte).

I’ve heard it described as “the number of addresses per square meter of the earth’s surface” and “like grains of sand”, but these examples weren’t working for me. For me, what works is the fact that 128 bits will give us 300 Trillion, Trillion, Trillion addresses – I repeat, 300 Trillion… Trillion… Trillion… addresses.

IPv6 – Introduction

Uncategorized No Comments »

“It was horribly like a spring flood; first a trickle, then a stream, then a torrent, then simply the grim struggle to keep from going under and…”
-Sarah Monette, Elizabeth Bear – 2007

IPv6 is coming!  IPv6 is coming!

It’s time to stop thinking that it’s a future thing – that it’s never going to happen.

You don’t want to be standing flat footed when it (the torrent) hits you in the back of the head.

06/23/08; Vol. 27, No. 15  “Agencies have until June 30 2008 to meet the Office of Management and Budget’s mandate to get their network backbones ready to handle IPv6 traffic. The Social Security Administration achieved that goal six months ago.”

09/19/08 — 11:11 AM The U.S. Postal Service plans to deploy an IPv6-capable video surveillance system to 40,000 postal sites across the country

“…For instance, just this past September, Arbor Networks placed worldwide use of IPv6 at only 0.0026 per cent, but according to the Google study, the percentage of computers using IPv6 access to the internet grew steadily from 0.192 per cent in August up to 0.238 per cent in October.”–/111963

First a trickle…
IPv6 is coming – better said as IP vee 6, not IP version 6…

Both Microsoft’s Vista and Server 2008 have IPv6 installed and running out of the box. These systems can automatically discover and start using a 6to4 router – to use IPv6 on the Internet. These systems are also automatically seeking an ISATAP router to facilitate IPv4 and IPv6 coexistence on your internal networks.

Starting next year, I’ll be proving a blog version of a crash course on IPv6.  You’ll need this before you get that first IPv6 user call (e.g., What does FE80 mean?) or before your your manager announces that the ISATAP router (?) needs to be configured.

Global Name Zone in new DNS

Uncategorized No Comments »

I really liked this new feature GLOBALNAME ZONE in new DNS in server 2008 and thought it was particularly neat.

To help customers migrate to DNS for all name resolution, the DNS Server role in Windows Server 2008 supports a special GlobalNames Zone (also known as GNZ) feature. Some customers in particular require the ability to have the static, global records with single-label names that WINS currently provides. These single-label names typically refer to records for important, well-known and widely-used servers for the company, servers that are already assigned static IP addresses and are currently managed by IT-administrators using WINS. GNZ is designed to enable the resolution of these single-label, static, global names for servers using DNS.

GNZ is intended to aid the retirement of WINS, and it’s worth noting that it is not a replacement for WINS. In GNZ, after the creation and enabling of the GlobalNames zone, the administrators must manually create, add, edit and, if required – delete, name records from that zone. GNZ does not support dynamic updates.

So lets start taking the advantage of this feature.

New Features of Server 2008: Server Manager

Uncategorized No Comments »

Hey Guys!  So continuing with our research on Windows Server 2008, I wanted to highlight just a few more features:
I have experienced that Server Manager in Windows Server 2008 provides a single source for managing a server’s identity and system information.

Server Manager makes server administration more efficient by allowing administrators to do the following by using a single tool:
-    View and make changes to server roles and features installed on the server.
-    Perform management tasks associated with the operational life cycle of the server, such as starting or stopping services, and managing local user accounts.
-    Perform management tasks associated with the operational life cycle of roles installed on the server.
-    Determine server status, identify critical events, and analyze and troubleshoot configuration issues or failures.
-    Install or remove roles, role services, and features by using a Windows command line.

We should also get used to the new terms like roles and features.  Roles could add more functionality like DHCP, DNS, whereas Features can augment the functionality of installed roles, like Failover Clustering, and Group Policy Management.

Given the confusion, I will attempt to clarify, at the risk of adding more confusion.

Uncategorized No Comments »

October 13, 2008 (Computerworld) Microsoft Corp. announced today that the code name for its next operating system, Windows 7, will be the product’s official name.

Now that the official name for “Windows 7” is “Windows 7”, everyone is officially confused.

I’ve never been a big fan of using the year as part of the product name, e.g. Windows 2000 Server, Windows Server 2003, and Windows Server 2008, because it hides some useful information. I prefer the use of versions number, which helps clarify when it’s a major (architecture change) or minor (cleanup) release.

BTW, did you notice the switch from “2000 Server” to “Server 2003”? This happened when the 2000 Professional upgrade (the desktop release) became Windows XP – giving us the XP/2K3 pair, client/server pair.

Some say that the XP is for “eXPerience”. I’m of the opinion that XP is for the Greek letters “chi” and “rho” – Microsoft had been talking about its Cairo vision for years and this was pretty close.

Sorry about the distraction, I need to get back to the “version numbers”. Well, maybe one more side trip. What does the NT in Windows NT stand for? You and most people would probably respond: “New Technology” and I would agree if you wanted the ‘much later’ marketing answer. But actually it stands for “N-Ten”, code-name for the Intel i860, the initial development target for Microsoft’s new NT O/S. And the sticking of “Windows” on top of NT was a pleasant and fortuitous happenstance.

The first version of “Windows NT” was v3.1, not v1.0. Even though it was v1 of the O/S, Windows had progressed up to v3.1, after its v1 and v2. This gave us the happenstance of a popular interface (Windows) with a major new O/S (NT).

What followed was NT 3.5, NT 3.51, and finally NT 4 (the “To SUR, with Love” release for James Bond fans) with SUR being “shell update release” – major new interface, sometimes called the Cairo interface, first seen on Windows 95, but developed for NT. This is when we started “right-clicking”.

NT 4 was then, actually the fourth release and was reaching a normal maturity level for O/Ses – stable and commercially successful. But a number of companies, especially the larger companies, had shared with Microsoft, that no matter how good the O/S, they would not embrace a solution heavy in proprietary technologies (e.g. NTLM, NetBIOS).

Microsoft got the message and started a major, (new architecture) multi-year project to develop Windows NT 5 – note the major version number. NT 5 replaced, with considerable engineering effort, the proprietary parts with preferred industry standards (e.g. Kerberos, DNS, LDAP), while still maintaining backward compatibility.

At the very last minute, the marketing people decided to rename the O/S to Windows 2000. I can understand how everyone wanted something called 2000, at the time. But without the version number, and even worse, dropping the use of NT, the confusion began. The one advantage was that it faked the Windows 98 people into upgrading to a major new and completely different O/S (NT vs. DOS), while thinking that it was just a simple upgrade. The 98 people had been intimidated by the exotic NT.

With a major release (NT 5), there’s always a lot of cleanup to do. So engineering started immediately on a cleanup of the desktop (client) piece and soon was able to release NT 5.1 (minor release), known to the world as Windows XP.

On any of the O/S, you can type, at the command prompt, either “ver” or “winver” and the O/S will disclose the major.minor information plus any SP (Service Pack) details. As another happenstance, by using 95, 98, 98SE, ME, then XP as product names, Microsoft was able to move (fake out) the last of the installed base from the older DOS to a modern O/S, NT.

Next, engineering went to work on the Server half of the O/S and shipped Windows Server 2003, which is NT 5.2 (minor release), giving us the matched (cleaned up) pair of XP/2K3 – minor releases, “clean-ups” of the Windows 2000 (NT 5) major release.

After addressing the industry standards, Microsoft then turned its attention to “the security issue”. Microsoft had decided that they wanted to be known, in the industry, “for security”, not for a lack of security. XP SP2/2K3 SP1 is the matched security pair – the limit of what could be done without a major architectural release, NT 6.

NT 6 “Longhorn” was driven by major changes in the architecture to improve security (e.g. UAC, Service SIDs). Vista SP1/ Server 2008 is the matched security pair. Why would one upgrade to Longhorn (NT 6)? Answer: Security. IE7 has a unique “Protected Mode” on NT 6.

If you follow the “hidden” versions for Exchange and IIS, you also learn a lot about the major vs. minor releases. IIS is v6 on Server 2003, not v5.2 – critical information, if you want security for your web sites.
So, NT, the kernel (looking at the business side of things) has gone from 3.1 to 3.5, 3.51, 4.0, 5.0, 5.1, 5.2, 6.0, and then 6.1.

And Windows, the interface (looking at the consumer side of things) has gone from Windows 1.0 to 2.0, 3.0, 4.0 (9x/SE/ME), 5.0 (XP), 6.0 (Vista), and finally Windows 7.

But the attempt to clarify, adds a little confusion, in the fact that it looks like Windows 7 will still be using the “Longhorn”, NT 6.1 kernel.

If we’re just counting Windows releases, then Windows 7 works.

But the next release of the Server will probably, correctly be called Windows Server 2008 R2 – not a SP, not an architectural change, just a mid-life refresh, with new “roles/features”, what we saw with Server 2003 R2 (the first R2). This whole R2 thing is another discussion.

So Windows 7 could have been Vista R2. But with all the engineering focus on the interface, can you say “touch”, Windows 7 works for me. And, I’m hesitant to add, that it would allow the name Vista to slide into history.

If you’re not confused yet, Windows Server 2008 comes out of the box with a SP1 pre-applied (non optional). So when they ship its “first” SP, it will be called???? I’m going to stop now.

Live Chat