Boris Voronin from EMC’s Proven Solutions team talked with me about how EMC is the first company to leverage Microsoft’s Synchronous Replication API for Exchange 2010. This product Replication Enabler for Exchange 2010 plugs into the DAG framework but replaces native log-shipping with synchronous block-based replication. Very cool for customers who have a need for higher RTO/RPO and would like to leverage existing SAN technologies for Exchange replication. Recorded at Microsoft TechEd 2010.
Like a storm, TechEd 2010 came and went pretty quickly and brought with it a thundering roar of activity from all of the vendors and partners and gurus all fighting to highlight their message through the creative use of Twitter, $2 bills, Ducati’s, Treasure Chests, and attractive people. Here’s a nice collection of photo’s from somebody who goes by Sektormedia.
Here were some highlights of the New Orleans event for me.
Compared to last year’s Los Angeles location and the ole conference standby Las Vegas, I think New Orleans is a great location for the events. It’s got a huge (albeit aging) conference center and loads of places to see and things to do. Unlike Vegas, it has some fascinating history and a culture and lots of great music. I haven’t been to the city in about 10 years, and I’m amazed that each and every night, people on Bourbon Street party like Lindsay Lohan. It’s really a surreal street, almost like a different planet, and when you leave it and jump on Canal Street you return back to a somewhat normal city with somewhat normal smells. In the French quarter, the food was great. I was treated to a fine meal at Cochon and had a nice meal with a large group at Mr. B’s where I had my first profiteroles. Wow they were good.
The locals seemed to have mostly recovered from Katrina (at least the ones I met) and the latest disaster (oil) didn’t seem to faze them until you ask them where to get decent seafood. One cab driver said he wouldn’t dare eat the seafood these days… too bad. New Orleans really needs a break.
We had a great theme that reflected the larger EMC messaging around Private Cloud and had lots of traffic and lots of questions… what is the private cloud? Tell me about VPLEX… 90% of the time I was describing the Private Cloud and VPLEX. There was a ton of traffic for all other areas of the booth as well.
- Txomin Barturen and the Hyper-V Scalability Session – huge Hyper-V clusters
- Txomin Barturen with Elden Christenson and the Geographically Dispersed Clustering session – automatic multi-site failovers
- Brian Cote and the Exchange Archiving Panel Discussion – SourceOne is designed from ground up for Exchange and SharePoin
- Dustin Smith and the Exchange Masters Panel Discussion – EMC knows Exchange
Things I learned:
The things you can do with SQL and a few developers is pretty amazing. A couple of us played an XBOX that was connected to SQL Report Services and PerformancePoint and while we raced around the track, spectators could watch our acceleration, braking, average speed and about 10 other custom metrics – all plotted real-time on a big HD display. Although we had fun racing around, it made you think about the power of real-time analytics that could be applied to various other industries. A great demo, although the feel of the steering wheel reminded me of a 1985 CJ5 jeep I used to drive around in high school. Anytime you hit a bump, there was about 30 degrees of freedom to the left and right.
Along the same lines in terms of visual eye-candy, an experienced SharePoint admin also gave an amazing session about how she put together a really consolidated view of a 19 prison in Alabama (Folsom County Prison was included in there for you Johnny Cash fans). She used every possible connection point that you could imagine into SharePoint:
- Outlook email invitations were tracked and meeting attendees were recorded in a central location
- Agenda items from each meeting were tracked and saved for years
- Dashboards were created to highlight the most significant items for a variety of roles (staff, warden, analytics, just to name a few).
- And the whole thing looked great – way better than just a place for dumping files, it was a place where you were given information that matter to you and she made it easy for anyone to get the reports, metrics, and data they might need to do their job. Teams don’t make portals like that, that is the power of one well-trained person being empowered to deliver results.
Great plug for EMC’s Replication Enabler for Exchange
Ross Smith spoke in his session about EMC being the first and only company to provide Exchange 2010 Third Party Synchronous Replication API Support. We’ve done it through a product called the Replication Enabler for Exchange .. it’s a free plug-in for MirrorView or RecoverPoint. What is the point? It replaces native DAG log-shipping (asynchronous, LAN-based) with synchronous block-based storage replication (SAN-based). This allows customers a better option for Exchange 2010 DR if they already have an existing investment in SAN, or who would like to receive the benefits of synchronous replication (zero data loss). Also many customers want to have one DR product/strategy for all of their applications, and not one DR product/strategy per application. It’s all about choices. If you are a Symmetrix customer we have a product called AutoStart which has been updated to provide the same functionality.
Odd Stuff: Exchange Team’s Creating New Definition of the Term Thin Provisioning
There was a session that talked about Exchange 2010 and how they not recommending Thin Provisioning – which they stated is difficult and operationally complex. I’ve talked about thin provisioning (EMC’s term for this is virtual provisioning) and the benefit of having wide-striped pools of disk – which can greatly simplify configurations and add a huge value where storage needs will grow to unpredictably large amounts over time.
A few excellent use cases for thin provisioning:
- Exchange Mailbox databases
- SharePoint Content databases
- Exchange Archive databases (separating into different DB is new in Exchange 2010 SP1)
So am I wrong, where did I fall off track?? Why would I recommend thin when the Exchange team does not??
Well it turns out we are talking about two different things:
Exchange team definition:
You can start with less mailbox servers and fewer databases than you actually need! (they called this “thin provisioning”). Then, over time, you can add more servers, and bigger databases, and then move mailboxes (move mailboxes!) to grow the storage to it’s full capacity. huh? Adding Less Mailboxes Now and Adding More Later is not thin provisioning. It’s called “normal”. Come on guys, you aren’t doing the industry any favors by making up definitions for terms that already have a definition.
EMC definition for thin provisioning:
Create a large pool of disk for use by your applications. Start it small and let it grow over time. When the space gets used up, expand it online, non-disruptively. Repeat as necessary.
There’s always something eye-catching and fun at TechEd, and I’m always happy to go – and always happy to point out the differences in the messaging between our company and the other schools of thought.
Longest title ever? Thankfully I abbreviated SCVMM down from System Center Virtual Machine Manager. Anyway…
EMC will be there to support them in many cities including Baltimore, NYC, Irvine, Raleigh, St Louis (to name a few).
I was asked to see if we could put together a quick demo showcasing some of the cool stuff we could do, and we hooked it up FAST.
My colleague Ryan Kucera and I worked together to put a quick little proof of concept together showing a combination of dynamic storage and server load balancing. In little over a week (just before his next proof of concept build-out), we were able to crank out a demo that showcases:
- System Center Virtual Machine Manager R2 (beta)
- Hyper-V R2 Live Migration (not released yet)
- Exchange 2010 (not released yet)
- SQL 2008 R2 (not released yet)
- CLARiiON Virtual Provisioning (creation of thin LUNs)
- Storage IOPS thresholds (Navsphere Quality of Service Manager aka NQM)
The setup of the demo was this:
You’re setting up your virtual servers on Hyper-V servers and you’re moving stuff around pretty quickly… You place two busy VM’s on the same host. Performance is bad. You need to move the VM’s without downtime – we use Windows 2008 R2 Live Migration to show this. Then you notice because we are using CLARiiON Virtual Provisioning and Thin LUNs for simplified management, we have multiple heavily utilized LUNs for different VM’s that are competing with each other on the same set of disks. No problem. NQM gives you the ability to be able to place a threshold on LUN’s (like 500 IOPS max for SQL 2008 R2 in the video) and let others (like a standalone Exchange 2010 VM in the video) have more IOPS to service more requests.
Too many people don’t know most EMC storage devices can do this (in both physical and virtual environments).
But now you do.
(looking for higher resolution on the video – click here)
Interesting article by Paul Robichaux over here on Windows IT Pro as he ponders a similar question I’ve been hearing a lot lately.
If I’m on Exchange 2003, do I upgrade to Exchange 2007 and then Exchange 2010 – or do I wait and go straight to Exchange 2010?
poll results here:
As always, this stuff (polls/surveys/etc) is never scientific and it could be as little as 100 people who responded to this, however the results are worth considering – especially if your company has products that are designed to work with Exchange 2007 today. Pretty soon your customers might start knocking on your door for the 2010 versions.
The tech industry is filled with “analysts” who will receive a buck and write exactly what a vendors asks of them.
In this installment, Symantec went to the “Tolly Group” and got them to blast EMC’s Networker for being slower than Symantec’s new feature for backing up and extracting single mail items, called GRT for short. In their graphical depictions, they show how apples compare to watermelons.
It leaves one to wonder how much Symantec spent on that piece of marketing, and time will tell if users believe those claims.
While Symantec offers what appears to be an interesting way to do a full backup and a brick level backup in one step, this approach is unsupported by Microsoft. Hmmmm, support is pretty important when you are talking about backups of a mission critical application, yes?
EMC and the Networker team do not recommend brick level backups, which are long, painful, direct MAPI scans and pulls. Instead, they recommend using the Recovery Storage Group, as the Microsoft supported and recommended way to achieve single item restores with newer versions of Exchange such as Exchange 2007. The Microsoft Exchange team in general is moving away from streaming backups altogether, so VSS (Volume ShadowCopy Services) and RSG recoveries are really the direction most companies will be taking – and the EMC NetWorker mailbox/message recovery solution does not have any of the long list of limitations offered in the Symantec solution.
And a really, really simple way to handle those quick “I need a piece of email back” requests is to turn up Deleted Items Retention to a larger amount. This gives the user the ability to find their own mails that may have accidentally deleted without any administrator time lost.
A few last data points worth mentioning:
- Exchange GRT Recoveries are only is valid for backups to disk folders.
- Exchange GRT Recoveries are a single, TWO PASS backup. Granular backups/recoveries take longer to allow NetBackup or Backup Exec to index mailbox/public folder information
- Support for Exchange Full Backup only
- Microsoft Services for NFS must be installed on the Exchange server
- Follows the best practices of the Microsoft Exchange Team for mailbox and message recoveries
- Allows mailbox and message recoveries from backups made to disk, tape or VTL
- Supports mailbox and message recoveries from source-based and target deduplication backups
- Can recover a large Exchange mailbox from the RSG in less than 8 minutes and an Exchange mail message from RSG in less than a minute
I couldn’t figure out how to embed this into my blog, but this is a great, short video which shows how EMC is working with Microsoft to virtualize applications like Exchange, SQL, and SharePoint. It’s only ten minutes long and showcases one of EMC’s great technologists, Brian Martin, as he speaks with Microsoft’s Jim Schwartz.
For EMC World 2009, I’ve got two presentations that I’m putting the finishing touches on:
1. Exchange Disaster Recovery: Good, Better, Best
This will highlight key differences in the styles of replication available to an Exchange environment: host-based, appliance-based, or array based. I’ve got some very interesting data comparing the built-in options of Exchange 2007 CCR and SCR versus EMC RecoverPoint and our array based offerings SRDF and MirrorView.
2. Microsoft Exchange 2007 Storage Best Practices for SAN and DAS Deployments
This session will discuss our latest best practices and discuss how EMC can offer low-cost solutions that meet the budget as well as mission critical deployments for businesses that truly consider Exchange to be a mission critical application.
On a related note, a recent customer asked “what do you consider to be mission critical?” My answer was “if your employees would flip if Exchange was down for a few minutes, then it’s mission critical. If you use Exchange for just about everything, it’s probably mission critical. If you are sending alerts that mark specific business events and provide notifications to line-of-business managers, then it’s probably mission critical. If you lost email data, and you’d fear your job – it just might be mission critical.” Sorry for sounding so Jeff Foxworthy.
Here’s some documents from various departments within EMC around the new Symmetrix V-Max and Microsoft Applications.
If you are implementing Exchange, SQL, or SharePoint in your environment today, I highly suggest you take a look (and read) the following:
EMC Proven Solutions
- Reference Architecture for V-Max with SQL, SharePoint, and Exchange
- Whitepaper for V-Max with SQL, SharePoint, and Exchange
EMC Partner Engineering Whitepapers
Today, Microsoft has released a public beta of Exchange 2010 and has affirmed the name publically (although the name slipped onto the web via Paul Thurrott SuperSite months ago). Here’s the new Microsoft Exchange 2010 Homepage.
So far, the major themes being highlighted via Flash animation include:
- Unified messaging
- Exchange Online
- Archiving and Retention
Download it here:
Consider EmailXtender to be EMC’s first generation archiving product.
SourceOne is EMC’s Archiving 2.0 product – completely re-written to match today’s needs.
The most significant and unique difference to me is the ability to assign roles to specific jobs within their archiving framework. In this sense, EMC is releasing a significant archiving product which looks a lot like some of Microsoft’s popular enterprise applications out there.
In Exchange you have Hub, CAS, and Mailbox roles… in SharePoint you have web front ends, app servers, and database roles. These roles can live on the same server, but in large environments people might prefer to break out individual components to allow for max performance, scalability, and reliability.
Like Exchange and SharePoint – SourceOne effectively breaks out all of the unique roles within the archiving framework to allow for flexible re-distribution of resources as your requirements change. “Resources” in this case includes input sources, workers, archive, and the master server. They can all be on the same server or broken out over time onto dedicate physical or virtual servers.
Over time the amount of data you will archive could increase – you can scale a component rather than re-architecting the whole system. As we add other archiving inputs (perhaps SharePoint) you will need to make sure there are enough workers and archive servers for the new data.
And you can break out components temporarily. Let’s say you want to pull in all of your PST files across the company… you’d provision a few servers for the task at hand during off hours, then re-purpose them to something else after the ingestion is complete.
This is a big announcement and it makes me happy we didn’t do it yesterday
Interesting links to check out
Even if you don’t take the class, you can learn a little about how the product works by reviewing this.
Coverage from other EMC Bloggers
Chuck gives his high level view of how this product enables enterprise-wide ILM and how ILM benefits organizations.
Zilla talks about how email archiving products really haven’t caught up with the rising tidal wave of email within a company.
I’ve been traveling a bit lately and spending a lot of time working with customers and presentation materials… so I haven’t had too much time to come up with just one big topic. Instead, I will offer you a few snack-sized morsels. Enjoy.
Cisco and Unified Computing and … Microsoft?
Seems like everyone needs to cover this in their blog, and since my angle is typically Microsoft oriented, it’s interesting to note Cisco is also partnering with Microsoft on their Unified Computing System strategy. The Cisco strategy – if you haven’t heard yet – is this: “pretty soon we’ll be selling servers, and right now we’re lining up as many big players as we can get.”
“The Cisco agreement with Microsoft was of a more limited scope, with Cisco agreeing to resell and support Windows Server 2008 with Hyper-V technology, Windows Server 2003, and Microsoft SQL Server 2008. Microsoft also said that it would work closely with Cisco to ensure that the Microsoft System Center family of products integrates with the Unified Computing System. (Windows IT Pro)
Dugie’s Pensieve Provides Microsoft Virtualization Guidance
EMC is mentioned in both the GUIDANCE summaries provided by Dugie. Virtualizing SQL calls out the storage utility I blogged about not too long ago with Symmetrix DMX4 (with virtual provisioning) as the base technology. Virtualizing SharePoint points to Microsoft’s own guidance on Virtualized MOSS and also makes reference to a friends blog who writes about how EMC’s Proven Solutions Team proved that virtualizing SharePoint can save 70% of the power resources compared to traditional physical deployments.
Microsoft SCVMM 2008 R2 Beta Now Available
If you are still doubtful that Microsoft is going to release a Live Migration (VMotion-like) functionality, you’re mistaken. It’s here now (in beta form). More details and how to get it HERE.
Internet Explorer 8 Ships.
If you are not a Chrome fanatic or a Firefoxer, you can start by downloading IE 8 here. Compatibility mode will help you adjust to pages that aren’t yet ready for IE8. An expanded index of add-ins (popularized from Firefox) allow you to download a fairly slim base install, then customize over time with web developer tools, search engine optimization plug-in, and a lot more.
Microsoft Research’s WorldWide Telescope goes public.
“you can bring up the Whirlpool Galaxy (M51) in optical wavelengths, load in Spitzer’s infrared view, and then use a slider button to morph one view into the other. The dark, dusty swirls in the optical view turn out to be exactly where the infrared emissions are the brightest.”
Cool! Download it here.
Azure Cloud Gets a Refresh
As you know, Microsoft’s V1 products aren’t the best. But look out for V2, V3, and especially V4. Microsoft is now on V2 (or Beta 2) of their Azure layer of cloud services that sit atop their “Red Dog” cloud operating system announced last October – publicized in the media and in magazines I like including Wired. This means it’s getting closer and closer to showtime.
Microsoft announced last week that it is also reworking the SQL Data Services (SDS) component of Azure to make it more like a hosted relational database. They want to make sure there’s an alternative to MySQL in the cloud.
More information on Exchange 14
Well, not too much new information, but enough to keep the herds at bay. Exchange “14″ OWA will not only support IE, but also Firefox and Safari. Wonder if Chrome will come next??
Next time, probably one topic, more thought. Thank you for reading Power Windows. Subscribe via RSS feed if you’d like to stay tuned to what’s happening around Microsoft in the Enterprise.
As you probably know, server virtualization helps you reduce costs on servers, power, cooling, and management and opens up some new options for backup, HA, DR, and adds a new level of flexibility.
Since Microsoft now supports Exchange, SQL, and SharePoint virtualized environments – whether it be on Hyper-V or VMware through their Server Virtualization Validation Program (SVVP) – it makes sense for every company out there to at least consider it. EMC is putting on a seminar series that aims to help customers understand the benefits and challenges of this new world.
At this free seminar you will learn:
- Why it’s OK to virtualize Microsoft applications.
- Why some applications actually run better in virtualized environments.
- How these applications perform and scale in a virtual world.
- How we’ve proven each solution we’re talking about, and documented detailed cookbooks and Reference Architectures for customers.
- How to approach backup in the virtual world.
- How to approach DR for Exchange/SQL/MOSS in the virtual world.
We’ve scheduled a few cities so far, but we’re always interested in talking to more customers and prospects, so shoot us an email and we’ll let you know what we can do to address your questions related to virtualizing Microsoft apps. I’ve been building some new powerpoints and demos for this roadshow and I can’t wait to finally show this stuff to some people.
Email us for more information:
MSSeminar@emc.com (there are two real friendly people – Cathy and Maureen – behind this list)
I met with a company yesterday that didn’t understand what tiering their Exchange users would do…
In their company:
- Everyone had the same mailbox limits – some mailboxes up to 7GB!
- Anyone could get a Blackberry (good for users, 3.64 IOPS multiplier for the Exchange server).
- Lots of Entourage (with Exchange 2003 it’s known to be “special” on IOPS).
- They didn’t offer single item restore because they don’t do brick-level backups or use Replication Manager.
- Recovery Times were “best effort” and often beyond 24 hours (they admitted this is bad, this was the point of the meeting).
In speaking with a lot of customers each week, I find that the above is pretty common. A lot of companies aren’t necessarily optimizing their environment.
A really nice, simple way to balance user needs with business objectives is to gather round the table and jot out a list of requirements for various parts of your business.
Believe me, I know there are all sorts of hurdles like political pressures, top down management styles, and weird people you want to avoid but it doesn’t seem too difficult for most companies to at least organize a single meeting where storage / network / server / Exchange / virtualization teams all sit in the same room and discuss requirements with the goal of finishing a matrix of offerings.
Need help? Maybe ask a business person in your company. Or ask your favorite vendor to help. Maybe ask EMC We actually have all sorts of quick services to map all of this stuff out for you.
If you’re interested, here’s a REALLY simplified version to borrow from.
Someone told me I should put my YouTube video on here. Over 5000 views… not bad.
I wish you luck trying to read my chicken scratch handwriting
Compared to past versions, the architecture of Exchange 2007 is split up into various smaller roles, each of which can run on a single physical box or a single virtual machine atop a larger physical box. Also, did you know there is no way to do an “in-place upgrade” to Exchange 2007 from a previous version. So, you will need to buy new 64 bit hardware to support the new infrastructure, and you will be building this environment next to your existing Exchange 2000/2003 production deployment. And then you will move mailboxes over… one server at a time. And… almost forgot… the economy sucks, so people are looking for ways to save costs.
More than ever before, I think now is a great time to make the move to a virtualized environment for Exchange 2007. Really, there are no major downsides anymore.
Besides taking the time to go out and learn about virtualization, what reasons is there for not virtualizing Exchange?
(this is a scheduled post. I’m actually on vacation somewhere in the middle of Florida. It’s got a monorail and rides and princesses(!) for the little ones. Yeah, that place. Anyway, if you comment, don’t be offended if it takes me awhile to get back to you, I could be in a 2 hour long line to see Cinderella.)
Thus blurring the line between what TAP program participants (such as EMC Corp) can say to their customers and what Microsoft is leaking publicly…
My rule of thumb?
If it’s not on the web somewhere else, it’s probably confidential.
Exchange Team Blog: A New E14 Video.
TechCrunch: “70,000% Increase In Beta Testing For Upcoming Microsoft Exchange Release”
Let’s hope they back off the whole direct attached storage nonsense… SAN’s still make the most sense for any mission critical application in the enterprise. DAS is simply a choice for those who want a lower acquisition cost.
[Looking for Exchange 14 or Exchange 2010 information? Click here]
The Microsoft Exchange team was pretty quiet in 2008, according to the most commonly available chart indicating Microsoft Exchange versions and their release dates:
Oh yeah, this leaves out 12 Rollups. 7 for Exchange Server RTM and 5 for Exchange Server 2007 SP1.
So, maybe it’s about time for Exchange 14 aka Exchange 2010 to start to reveal itself?
If the name Exchange 2010 is news to you, the name is starting to make appearances around the web, although I think the news was broken by Old Reliable, the Windows SuperSite by Paul Thorrutt.
What’s in Exchange 2010?
Although EMC is officially part of the “Exchange 14″ TAP (Technology Adoption Program), I would be slapped silly by a big guy in a blue shirt if I revealed some of the documents I’ve seen.
However, let’s consolidate some public speculation and discussion I found on the web:
- Microsoft Consultant Doug, who writes Doug’s blog, offered a glimpse into what he thinks is coming in the next release of Microsoft Exchange Server which includes more copies of your databases, cheaper storage, no RAID (?), no backup (?), virtualisation, and CAS for all clients. [Update: since my finding and promotion of this site, the article is surprisingly not available anymore to the public but thanks to Web Caching, here it is. (looks like this is gone now, I do have a screenshot of the original material)
- SearchWinIt/TechTarget writer Margie Semilof revives rumors of Exchange moving to a SQL-based data engine.
[Update 2: After the publication of this post, which was made with all publically available information, Microsoft released a video (without much meat) about Exchange 14/2010. Were the two things related? All I can say is perhaps…
The Exchange messaging continues from Exchange 2007, and customers are still being advised to make bad decisions (which is to run a large Exchange environment on cheap racks of disk). The customers I am working with are still finding value in their SAN’s and here’s some best practices based on the advice of some of our early adopters of Exchange 2010. Most companies understand a SAN is a great way to lower the long-term operational cost of Exchange, but how about lowering the purchase price?
Are you looking to lower acquisition cost of your Exchange infrastructure?
1. Virtualize it. You’ll get more out of the new hardware you’ll have to buy to support 64 bit Exchange.Today’s new hardware is so beefy by default, it is likely that Exchange might not be able to stress it completely (SQL on the other hand is a different topic altogether). Put multiple Exchange VM’s on it, and it’s a new game. It also provides the foundation for a Cloud strategy, if you are thinking about going that way.
2. Put it on SATA drives. Exchange 2010 disk activity is so low; you can run the entire environment on affordable (and big) SATA drives. You can also run a combination of Flash/FC/SATA and turn on EMC’s FAST and allow it to automatically adjust the workload depending on the performance measurements that FAST takes. You can move some workloads to the cloud and easily re-purpose the SATA drives for something else. With DAS, if you move to the cloud, you’re stuck with a lot of direct attached disk that you can’t easily re-use.
3. Turn on Thin Provisioning (EMC calls this Virtual Provisioning). You can start small and monitor mailbox performance and capacity. If necessary you can increase the space of the thin pool, dynamically. With DAS you are forced to overbuy storage and over-allocate a significant amount to provide for the large mailboxes that Microsoft wants you to have. With EMC’s Virtual (thin) Provisioning, we can start small and grow as you need it.
Example Config and Comparison:
- Exchange 2010 DAS, 5000 users with 5GB mailboxes = 25TB
- Exchange 2010 on EMC SAN, 5000 users with 1GB mailboxes = 5TB
So in this example, the DAS configuration requires you to buy 5x the storage, UP FRONT, and pay for the power, cooling, space, and the real estate for those disks up front. It’s most likely those disks don’t have a spin-down capability, so if you’re buying DAS for Exchange you’re buying a fleet of old gas guzzlers, turning them on, and leaving them in the driveway. If you buy the smart way, you’re getting something a lot more dynamic, that grows as you need it, something that’s a lot more efficient. Almost like hybrid cars. Except better.
Please watch my video and check out the picture – I am showing off how I can rub my belly and point to a whiteboard at the same time!
I just got back from Australia visiting several large EMC customers (saw Meg Ryan randomly when we were there). Even down under, the whole Exchange 2007 sponsored “SAN versus DAS” debate is still alive and kicking…
Logically speaking, isn’t it easier just to share and consolidate stuff?
Examples of “stuff”:
- Airplane video monitors. More monitors, more problems. Central systems are more efficient. First class passengers, sure they get all the video/movies/games they want. But they pay heavily for the convenience and additional control and of course for the additional maintenance.
- Bank accounts. Isn’t it easier to do your banking from one place instead of having different accounts for your stocks, mutual funds, savings, checking, money market, and mortgage loans? I’ve finally got my finances down to two places…
- Boats. Those who know say it’s much cheaper to chip in or rent than to own your own boat. Renting also lets you try out different “tiers” of boat – power, sail, speed, cruisers, etc.
- Libraries. “Networked libraries” where you reserve a book and have it sent to you are substantially better than searching or driving to each library one by one to find the book you want. I’m a beneficiary of a networked library (part of the minuteman library network). It’s awesome.
- Debt and student loans. You can group your liabilities and get a reduced rate from the hopefully not-in-default financing organization that wants the interest $ from you, or you the old way, you deal with each bill one by one.
- Companies/Government Agencies. Organizations (including companies/community groups/towns) are frequently consolidating to be able to provide shared services for one another. Hamilton and Wenham Massachusetts are actually discussing a merger to be able to lower costs by utilizing shared resources such as schools, rec centers, libraries, and potentially fire and police stations. They are already at a stage called regionalization (which is the direction towards consolidation).
- There’s probably more, but I am out of ideas at the moment.
Here’s another statement from the Town of West Boylston (again Massachusetts)
Regionalization allows communities to share administrative and program costs between two or more communities in order to increase or maintain the level of services within the participating communities. I often wondered why every community needed its own police department, its own fire department and its own library. I find it hard to believe that this system will exist 15-20 years from now as municipalities struggle every year to provide even basic services to its residents. One of the areas where Massachusetts has some success with regionalization is with regional school districts. Regional school districts allow communities to share costs and to provide services they could not provide if they were acting on their own.
More reading if this is not putting you to sleep:
Not long ago, I was giving a workshop to a few customers in Indianapolis. The point was to educate them on what EMC could do for Microsoft apps and databases.
At one point, during a short bio break, two guys were getting quite animated comparing the length of their nightly backup windows. One was a SQL DBA. The other was an Exchange admin.
I laughed pretty hard, not only because I live in a world where I am constantly surrounded by geeks, but because neither of them was accomplishing anything remotely remarkable with their backups. No dedupe. No VTL or EDL. Not even a proxy node to offload the production CPU cycles. Weak.
Seriously. If your backup window is in the hours, you are not only missing the boat, but you are also missing that time when you could be doing more useful stuff with your systems.
Time is important. I live outside Boston, and when I have to go to NYC, I sometimes drive – but rarely. I’d rather be on a plane or a train (big Accela fan) so I can be doing something (geeky) instead of being focused on the road for 3 and a half hours.
That freed up time allows me to read the paper, watch a movie, catch up on work, or make calls (not in the quiet car of course).
as quoted by Chris Capossela, Microsoft’s senior veep with the Information Worker Product Management group… it appears Microsoft is about ready to get serious about the web with a hybrid offline/online approach and SaaS offerings mixed with dedicated/shared/virtualized hosting models.
I frequently consult with customers who need to figure out how to better store, consolidate, protect, and manage their SQL databases, Exchange servers, and MOSS farms.
With Microsoft, most of the time I am simply in awe for their long-term ability to seed and harvest a market by slowing slipping innovations into gradually more expensive and higher-quality V2 and V3 products. SQL Server, Zune, SCOM, and Hyper-V are all illustrations of this. But sometimes you get a rogue product group within the company that acts without regards for their overall company strategy. And maybe that’s what happens when you have 65% market share.
I am speaking specifically about how people are being swayed away from using “expensive SAN hardware” for storing and protecting Exchange 2007 data. Not only do the advocates of this approach (which includes the Exchange product team) obscure long term costs by comparing pure “cheap storage” or “DAS” acquisition costs and ignore long-term cost of ownership, they also ignore:
- Management and simplicity of shared storage devices (one place to put stuff)
- Utilization benefits of shared storage device (one place storage means more likely to use it)
- Bandwidth requirements of the CCR/SCR replication scheme which can be 5 times larger than cached replication appliances (say EMC’s RecoverPoint)
- Exchange server virtualization – massive amounts of people are virtualizing their Exchange servers and placing data on a SAN to get the maximum benefit (virtual servers need virtualized/SAN storage to enable most of the advanced features).
Paul Galjan writes a very convincing blog post about dissonance within Microsoft, their latest anti-SAN calculator, and how this anti-SAN stance really backfires when most of the company is aggressively pursuing a very heavy virtualization strategy and taking a more balanced approach about which virtualization platform to use (like ESX being certified under SVVP). I travel around the country giving workshops on how to virtualize Exchange and it’s truly up to the customer to decide whether I talk about Hyper-V or VMware ESX as their hypervisor of choice. At the end of the day, the customer will guide that decision. Not me. I am not paid to push VMware.
If it’s really a cost issue, perhaps customers could chain a bunch of USB drives together and create a super-DAS configuration. But wait. That is pretty close to what they are saying, isn’t it? Get rid of the SAN, get a lot more servers and some cheap disk, and prepare to throw a lot of people at the new storage management problems that will be sure to take place because you aren’t sharing storage across servers, and you’re not sharing global hot spares with hundreds of disks in a system, or taking advantage of a nice big cache layer that smooths out most of the peaks during the Outlook users workday. And not taking advantage of high speed LAN-free backup snapshots or clones for recovery from corruption.
I’ve seen customers trade in SANs and array-based replication schemes for this new model, and only months later come back to an EMC SAN (and replication) saying “OK, you were right… Just don’t say I told you so.”
But it’s not like me to say that to my customers. I warn them up front.
Some interesting data I’ve been able to gather recently:
- This article talks about the long-term costs of the DAS approach (and did not focus on Exchange specific details such as bandwidth costs, latency, and ease of use).
- SearchStorage article about how a SAN can be more cost effective vs DAS in Exchange 2007 environment.
- And over at HP, we’ve found a friend in a strange place… as their server group is ecstatic about the additional servers required for the DAS solution, yet the EVA team is a little confused. A little Exchange lab testing confirms what I’m talking about (extra latency, difficulty, and management overhead).
- Here’s another reason why SCR could be a sketchy DR solution.
- Another one on the imperfections of this version 1.0 technology, on Microsoft TechNet.
- Oh, and don’t forget to run full backups on this to truncate logs.
- Even the Microsoft teams knows about the various other issues with this replication technology.
- And here’s me on YouTube, trying to sum up why a SAN makes sense for Exchange 2007 (just ignore the really bad hand writing).
For a while EMC has been running their own internal web events to allow the product teams and engineers to communicate directly with people who might become customers. Below is a list of several events that you might find interesting that are available OUTSIDE the firewall. Enjoy!
To further add to this, I’d like to comment on issues 2 and 3.
2. Storage hardware does not support SCSI 3 Persistent reservations
Well, it’s a good thing EMC is good to go here. When using EMC Symmetrix, you do have to enable SCSI-3 persistent reservations. To do this, you need to set a bit called the PER bit. More information available here.
3. Multi-Site configuration education – In Windows 2008, the Failover Cluster software allows for better support of multi-site configurations which allow the cluster nodes to be on different subnets (and support what they call an “or” feature… this subnet “or” this other subnet…) to reduce the need for configuring VLANs. The caveat here is the application support. EMC has had a products which support geographically dispersed clustering for quite some time and we’ve seen configured all types of VLAN configurations. What cannot be done is SQL based “or” functionality (because SQL doesn’t support yet). Exchange 2007 does support this functionality however.
EMC has long provided a variety of solutions for companies that want to keep data protected and available. For Microsoft applications like Exchange 2003 Exchange 2007, Exchange 2010 and certain SQL applications, this is no exception…
So, just about everyone who knows about EMC is probably aware of our CLARiiON replication technology MirrorView and the Symmetrix DMX replication technology SRDF. What I continue to find shocking, after traveling around talking to lots of customers, is that a lot of people don’t know about a simple mix between a cluster and our array-based replication. This clustering integration component is called a Cluster Enabler.
And this technology can be the pinnacle of Disaster Recovery: Zero Data Loss, and a Near-Zero Recovery Time.
Our customers say it best:
“The biggest advantage of SRDF/CE is that it enables us to treat our secondary site like a second room in a data center, which means we can put the active node at our main data center and the passive node at our recovery data center, and not have to buy additional hardware to support clustering.” – Christopher Palagano, OhioHealth
Update 1/6/09: here’s a link to the blog by one of EMC’s cluster gurus (and a Microsoft MVP), John Toner. It’s probably the only blog in the world that is focused exclusively on geographically dispersed clusters.
Update 7/15/10: forgot to mention we also have a version of a Cluster Enabler that works with EMC”s RecoverPoint product. EMC’s RecoverPoint is a powerful replication appliance that can journal data locally or remotely and enable rollback of data to an exact point in time hours, minutes, seconds, or milliseconds ago. More information on RecoverPoint/CE is available here.
Here’s a Proven Solution test we performed with RecoverPoint/CE. Our goal was to simply to reduce RTO and RPO.
Here’s what we achieved:
- Only 7 minutes were needed for a complete farm failover with this solution
- 4x times less network utilization with RecoverPoint bandwidth reduction
- Less than 1% overhead on CLARiiON storage processer with integrated RecoverPoint-CLARiiON splitter
- Microsoft Hyper-V used to consolidate 15 physical servers to 6 virtual servers