Planning a Data Center Migration: Industry Insights, Tips & Notes on India

Reasons for Data Center Migration

Business	Technology
-        Strategic Move -        Regulatory Requirements -        Merger/ Demerger	-        Consolidation of multiple Regional DCs -        Obsolete data center facilities -        Legacy IT infrastructure -        Fully utilized, out of capacity

Industry Insights on Data Center Migration

• By 2011 more than 70% of U.S. enterprise data centers will face tangible disruptions related to floor space, energy consumption and/or costs.
• 94% of IT departments are either considering or undertaking a server migration program. These initiatives are the most effective ways to lower operations costs (TOC) in a company’s data center and estimates that businesses can save up to 23% of annual costs with a properly executed plan for server migration.

Source: Gartner

• Estimates range from $300 – $400 per foot to move existing equipment and $750 – $1,000 per foot to implement new equipment

Source: Aperture

Planning for Migration

-        Decide to Build or Lease the new facility

-        Conduct a Site Suitability Analysis

-        Plan Thoroughly and Well in Advance

-        Minimize Complexity and Change

-        Be 100% Focused on Move Day

Tips for Planning Data Center Migration

• Even organizations that have fairly simple data center operations and/or very thorough documentation should begin planning well in advance of a move
• Companies should begin planning three to six months in advance  – DataMove
• Know the maximum tolerable downtime for an application
• Companies should allow plenty of lead time, especially when setting up high-bandwidth WAN connections from telco providers. – Networking Unlimited. Verizon requires a 90-day lead time for setting up WAN services.
• If applications are to be made available during the move, then redundancy and alternate site considerations are important. Eg. Company Website
• Power and Networking, are the 2 issues that surface most during a data center relocation. -Forsythe
• Consider the network latency for applications resulting due to the distance of the new facility from the existing users (or offices).
• Also, check if your equipments have ‘location-specific’ warranties from their manufacturers
• Divide the equipment to be moved into distinct move groups that correlate with a move sequence, placement in the new location and grouped according to priority or mission criticality. Applications that have a High Business Impact should normally be spread over multiple ‘move groups’ to lessen risk
• Evaluate which equipment will be retired and replaced at the new location, thus reducing the cost of relocation.
• Use the migration as an opportunity to implement Virtualization and other modern day data center technologies
• Consider engaging a third-party consulting firm to orchestrate the relocation, coordinate with the OEM vendors and the trucking companies, and to serve as relocation project manager.
• Establish backup contingency plans for each phase of the move and determine how critical applications will be supported if any phase of the move is delayed.

Typical Parameters for Application Migration

Source: EMC2

Signing Out Notes on India

• In India, infrastructure support like logistics & electricity may not be 100% reliable. So plan by keeping a buffer for the migration activities.
• Often, you might be required to have particular type of equipments (eg. Energy saving) when moving your facility to a Special Economic Zone (SEZ) in India.
• Often SEZs (like the one at Kharadi, Pune) wont allow your existing desktops, etc to be moved into the new premise. You will have to buy new equipments for the new premise. New equipments will be available at special rates in SEZs & many big equipment brands have tie-ups with the SEZs.
• Also, have proper tax papers (eg. any applicable import/excise taxes) done for other equipments when moving to such an area.
• SEZs also may also lay norms on activities that you can conduct from the premise.
• All in all, read the offer document carefully before investing!

Although these tips give an overall picture of the planning, do seek the help of professional consultants. Their proven expertise will help in hassle-free migration, which otherwise will be a pain area.

Do not forget to notify the users and other stakeholders of the migration activity & any resultant downtime.

Case Study – HSBC, India

Background

• Core Banking Data Center Based in Mumbai
• IT function in banking must be able to support the business with total integrity and without interruption

Reason for Migration

• Premium Office Space & Maintenance costs
• Consolidate physical servers through virtualization to save costs

Risks in Transiting existing equipment

• Longer transport  time
• Risk of damaging equipments/data during transport
• Intolerable Downtime

Approach Followed

• 3^rd Party experts: Pentagon Systems and Services
• Used PlateSpin Migrate from Novell
• ‘Decoupled’ workload from underlying server hardware and streamed them to virtual hosts
• Each server was moved during weekends on a four-week cycle
• The whole migration was completed within six months

Results

• Low-impact migration of servers
• No complexities of physical transportation
• Achieved inter-city migration of more than 100 servers across a distance of 387 miles
• Virtualized 40 servers, enabling hardware consolidation, increased flexibility and reduced costs
• Maintained complete system availability during working hours, minimizing user disruption

Read the entire case at Express Computer Online

Comparison of Routing Protocols

I Guess this should suffice for a fair understanding of the routing protocols mentioned.

 

TABLE 1	RIP v1	RIP v2	IGRP
Metric on which it is based	Hop Count	Hop Count	Bandwidth / Delay
Administrative Distance	120	120	100
Max Hop count	15	15	255
VLSM Support	NO	YES	NO
Uses Algorithm	Bellman-Ford	Bellman-Ford	Bellman-Ford
Content of Routing Updates	Full Table	Full Table	Full Table
Frequency of Routing Updates	30 seconds	30 seconds	90 seconds
Update Address	Broadcast	224.0.0.9	224.0.0.10
Protocol / Port	UDP 520	UDP 520	IP protocol 9
Convergence	Slow	Slow	Slow

 

TABLE 2	EIGRP	OSPF	BGP
Metric on which it is based	Bandwidth / Delay	Cost	Many (weight, path length, localpref, origin, age of the path, etc etc)
Administrative Distance	Internal – 90 External – 170	110	Internal – 200 External – 20
Max Hop count	224	-na-	-na-
VLSM Support	YES	YES	YES
Uses Algorithm	DUAL	Dijkstra	Best Path
Content of Routing Updates	Only Changes	Only Changes	Only Changes
Frequency of Routing Updates	Only when change occurs	Only when change occurs	Only when change occurs
Update Address	224.0.0.10	224.0.0.5	-na-
Protocol / Port	IP protocol 88	IP protocol 89	TCP 179
Convergence	Very Fast	Fast	Medium

 

The Technology Hype Cycle

Technologies come with a bang and disappear without any whereabouts. The Technology Hype Cycle explains the speculation and disillusionment surrounding technologies. In the later part of the article, I have tried to map it into another well known theory – Technology Adoption Life Cycle.

What is the Technology Hype Cycle?

The Technology Hype cycle is a trace of the hype associated with a technology through its life cycle. Every technology takes the market by storm as soon as it is launched – the ‘hype’. But over a period the users begin to find flaws (or eventually get bored). This forces the makers of the technology to work upon the flaws and/or introduce more features in their product. And this explains why technologies last for a considerable time.

Did we ever wonder that the photo-copying machine & our computers have also followed this graph!

This is the typical Technology Hype Cycle that tells about the way the technologies traverse the plot. It is released by Gartner every year & it is accompanied with a comprehensive analysis which is available to their premium members.

The following are the typical 5 phases of the Hype Cycle as per Gartner:

1. ”Technology Trigger”
This is when the product/technology is launched or other event that generates significant speculation in the people & press. Yes! The Hype, as we all know, starts before the product hits the shelves!!

2. ”Peak of Inflated Expectations”
According to me, this is the time when the technology is all over – tabloids, tweets, blogs, etc. and there is so much speculation massed up that everyone wishes to get a taste of it!

3. ”Trough of Disillusionment”
This is the phase when people find out any design/tech flaws or feel that its substitute was just better!

4. ”Slope of Enlightenment”
I am not sure if every technology will scale this slope! It’s upto the makers of the technology to brainstorm on WWW (What Went Wrong!) and come up with apt changes. Not everyone would want to again work on something and I guess there will be still some endorsers who would want to use their technology.

All in all this phase is where the makers of the technology how much to scale the slope and arrive at the…

5. ”Plateau of Productivity”
A technology reaches the “plateau of productivity” as people again come to know the benefits of it. It become widely demonstrated and accepted.The technology becomes stables and widely accepted here.

Untill…..again something disappoints the users leading it again to the Trough of Disillusionment…and the cycle continues!

Also note the expected time in years cited by Gartner for the technology to make it big. Of course, not all of the technologies go through these 5 phases. Some just fade awayand are never heard again; some stay around for years and don’t hit their hype peak! And some gain visibility at a steadier pace. But this theory of Gartner’s is still a useful visual metaphor, especially for high-profile technologies that do exhibit these traits.

I found a close resemblance of Gartner’s Hype cycle with the Technology Adoption Life Cycle.

While reading bout the Hype Cycle, I found a few points in common with another theory. The Technology Adoption Life Cycle takes off with the Innovators & Early Adopters who enthusiastically adopt any product/technology. Ofcourse, this theory later talks about crossing the ‘chasm’ (the big gap b/w Early Adopters & Early Majority) but it is interesting to know that the Hype is created by the Innovators & Early Adopter – Enthusiastic customers, Press, etc.

(I can’t offer credits to the creators of the Technology Adoption Cycle as there are too many of them who worked & re-worked. I got hold of this theory through Geoffrey A. Moore’s book – Crossing the Chasm. )

Where is Gartner’s Hype Cycle theory on the Hype Cycle? & did you check where my blog is in the cycle?!!!  

Thanks to my mentor for introducing such an interesting theory to me!

Reading Between the lines of The Indian IT Act, PART 1

This is first in a series of articles on the IT Act of India. I had this copy of the Indian IT Act for a while now & began going through it today. It doesn’t seem to be too exhaustive. I have read 9 pages from it. And I have the following for you:

Quick Facts about the Act:

• Actual name: The Information Technology Act, 2000
• Published as a part of ‘The Gazette of India, Extraordinary’ (Bharat Ka Rajpatr, Asadharan)
• Consent of the (then) President received for the IT Act : 9^th June, 2000
• It is recommended in the Act that if any state government wishes to make any amendments/changes, they should first refer the Model Law
• Interestingly, India’s IT Act is a derivative of Model Law on Electronic Commerce adopted by the United Nations Commission on International Trade Law (dated 30^th June 1997)

As you might be aware that all Laws/Acts have terms defined which are given a reference later. (& anything beyond the scope of the said terms becomes an exception!!). The definitions were a rerun of Basic Network Security class! It contains all the terms – digital signature, asymmetric keys, public/private keys, hashing et al!

Interestingly I came across some interesting definitions as well.

(Note: Below ‘ze’ is used in the Act as all the 26 alphabets were used and the list restarted from za, zb,…)

Section 2. Definitions

(ze) “secure system” means computer hardware, software, and procedure that—

(a) are reasonably secure from unauthorized access and misuse;

(b) provide a reasonable level of reliability and correct operation;

(c) are reasonably suited to performing the intended functions; and

(d) adhere to generally accepted security procedures;

Nowhere have they referred to what is reasonable!

(zh) “verify” in relation to a digital signature, electronic record or public key, with its grammatical variations and cognate expressions means to determine whether—

(a) the initial electronic record was affixed with the digital signature by the use of private key corresponding to the public key of the subscriber;

(b) the initial electronic record is retained intact or has been altered since such electronic record was so affixed with the digital signature.

Grammatical & Cognate Variations of a DS for verification? Interesting!

Section 9.  Sections 6,7 and 8 not to confer right to insist document should be accepted in electronic form.

Nothing contained in sections 6, 7 and 8 shall confer a right upon any person to insist that any Ministry or Department of the Central Government or the State Government or any authority or body established by or under any law or controlled or funded by the Central or State Government should accept, issue, create, retain and preserve any document in the form of electronic records or effect any monetary transaction in the electronic form.

Section 6 is about Use of electronic records and digital signatures in Government and its agencies.

7. Retention of electronic records.

8.  Publication of rule, regulation, etc., in Electronic Gazette.

I didn’t feel the need of quoting all these sections here as they are apt in the content. Point to see here is that the Govt. doesn’t take responsibility to retain & preserve any document any electronic form. There goes your Right to Information (RTI) Act for a toss!!

Calling all Lawyers and Cyber Crime Specialists!

(To be continued…)